Re: [opensuse] Beagle under 10.3 is really eating up my CPU
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Randall R Schulz schreef: On Sunday 13 January 2008 11:07, Randall R Schulz wrote: So a program like the Beagle indexer has to be smart enough about the load it offers to get its work done as quickly as possible without interfering with interactive use. By the way, I'm not saying I think Beagle does (or doesn't) need to do more or better along these lines. I don't have enough experience or information to have such an opinion. ... Randall Schulz As it got installed in the 11.0A0, i uninstalled it. But it got installed again... But there is changed something... It is very modest now.. So modest i even sometimes forget it is there! Maybe it read what was said about it in this list? :-))) So there is still hope..(one can only hope it stays that way..;) - -- Have a nice day, M9. Now, is the only time that exists. OS: Linux 2.6.24-rc6-git11-3-default x86_64 Huidige gebruiker: [EMAIL PROTECTED] Systeem: openSUSE 11.0 (x86_64) Alpha0 KDE: 3.5.8 release 31 -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.5 (GNU/Linux) Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org iD8DBQFHjHXUX5/X5X6LpDgRApoHAJ4hCTuna6Vw6YKiIqt/r3H5U3JXQACfWlTl RJbw0xTkTEb1pNIFVL09cEc= =msym -END PGP SIGNATURE- -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Anyone else seeing Beagle really kill performance? I have disabled it and my machine finally is perky, but every now and then, I find it in memory again. How do I arange it to chew up less memory and CPU or kill it once and for all? Usually this indicates you have a problematic file (usually its broken or corrupt) that causes the index helper to go into a loop while indexing. See http://beagle-project.org/Troubleshooting_CPU for instructions on how to report such a bug. A bit late to the discussion here... I also have to kill Beagle every time I do an install. I tried it again with the 10.3 install I did this weekend. It sucked up so much of my system resources that I could barely do anything else... this is on a *clean* default install (not an upgrade) on an AMD Athlon 64 X2 3800+ with 2Gb of RAM, a /home with basically no data, but about 1.2TB of data on other mount points. My CPU.. both cores.. were running about 99%. RAM was full, and swap was filling up as well. The whole computer was grinding to a halt. When I finally managed top open a terminal and run top... Beagle was there consuming 100% of everything it could. I left Beagle run for a while... an afternoon... and it never changed. Kept my CPU nice and toasty warm though. In the end I sopped the daemon, and removed every trace of Beagle I could find. The result... the computer is back to normal. The 10.3 install is noticeably faster than the previous 10.2 install (also without Beagle) and I'm happy.. .although a bit puzzled how it is that anyone finds Beagle useable. As a contrast, I can install the Google Desktop indexer (on the dual core system), and I never notice it is there. It indexes roughly the same scope of data (I think). It never runs so that I am aware it's indexing. My other apps carry on with no noticeable impact on performance. I see a few people here saying Beagle runs fine for them with no noticeable impact on performance... how? I've struggled with Beagle since it first appeared on the openSUSE scene. I have seen it's appalling impact on performance over several installs on several different hardware configurations. Not once have I seen it work in any measure that could be considered good. I will continue to try it out with each new install I do, but... i don't hold out a lot of hope. I've kind of lumped it in with zmd... another app that is on my search and destroy list for a new install. Once those two apps are gone from a default install the computer works great with openSUSE. C -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Monday 14 January 2008 02:22:18 am Clayton wrote: I see a few people here saying Beagle runs fine for them with no noticeable impact on performance... how? It seems that you monitor Beagle in a first time after installation. Though there is pops up note telling that computer will be slower in a first few minutes. Later on you shouldn't notice indexing. Though that it should be optional as you suggested in another post as the version is 'beagle-0.2.18-30' which by any interparetation of version string is early development. On the other hand, how many people will ever attempt to test software with so low version (except Linux users)? -- Regards, Rajko -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 The Monday 2008-01-14 at 09:22 +0100, Clayton wrote: Usually this indicates you have a problematic file (usually its broken or corrupt) that causes the index helper to go into a loop while indexing. See http://beagle-project.org/Troubleshooting_CPU for instructions on how to report such a bug. with basically no data, but about 1.2TB of data on other mount points. Which you could tell beagle not to index. That's a lot of data. My CPU.. both cores.. were running about 99%. RAM was full, and swap was filling up as well. which is probably indicative of a bug. I see a few people here saying Beagle runs fine for them with no noticeable impact on performance... how? See the first quote I left above. - -- Cheers, Carlos E. R. -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHiz9rtTMYHG2NR9URAsgeAJ4ko8h/Z5xzFJ0MDdxb1HfrDe+NDgCfURos llu11dAriA0ynsI+bww9sX8= =i+gd -END PGP SIGNATURE- -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
I see a few people here saying Beagle runs fine for them with no noticeable impact on performance... how? It seems that you monitor Beagle in a first time after installation. Though there is pops up note telling that computer will be slower in a first few minutes. Later on you shouldn't notice indexing. I let Beagle run longer than 24h on the dual core system. System response remained horrible. A friend installed 10.2 and then updated everything including Beagle... it ran for a couple of weeks with Beagle killing his system performance before he called and asked what was wrong. So, I am not talking 30 seconds of annoyance here... this is days of uptime on fast machines... and weeks on slower machines. version is 'beagle-0.2.18-30' which by any interparetation of version string is early development. On the other hand, how many people will ever attempt to test software with so low version (except Linux users)? Test by choice is a good thing... lots of us here install from Factory just to see what works. I have a VM I do that in all the time. Lots of things break and I have to roll back to a previous snapshot (which is why I like to use a VM instead of a native system) Setting it as part of the default install makes the new users test it as well. That isn't giving the new user a lot of choice. with basically no data, but about 1.2TB of data on other mount points. Which you could tell beagle not to index. That's a lot of data. True, but a significant portion of it is video. Mostly very large files eating up a lot of that diskspace not millions of small text files that need to be indexed. Indexing 2 or 3 hundred binary video files should not take that long. C. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Clayton wrote: Anyone else seeing Beagle really kill performance? I have disabled it and my machine finally is perky, but every now and then, I find it in memory again. How do I arange it to chew up less memory and CPU or kill it once and for all? Usually this indicates you have a problematic file (usually its broken or corrupt) that causes the index helper to go into a loop while indexing. See http://beagle-project.org/Troubleshooting_CPU for instructions on how to report such a bug. A bit late to the discussion here... I also have to kill Beagle every time I do an install. I tried it again with the 10.3 install I did this weekend. It sucked up so much of my system resources that I could barely do anything else... this is on a *clean* default install (not an upgrade) on an AMD Athlon 64 X2 3800+ with 2Gb of RAM, a /home with basically no data, but about 1.2TB of data on other mount points. My CPU.. both cores.. were running about 99%. RAM was full, and swap was filling up as well. The whole computer was grinding to a halt. When I finally managed top open a terminal and run top... Beagle was there consuming 100% of everything it could. I left Beagle run for a while... an afternoon... and it never changed. Kept my CPU nice and toasty warm though. In the end I sopped the daemon, and removed every trace of Beagle I could find. The result... the computer is back to normal. The 10.3 install is noticeably faster than the previous 10.2 install (also without Beagle) and I'm happy.. .although a bit puzzled how it is that anyone finds Beagle useable. This is typical behavior, and the developers have known about it for a long time, because there are written complaints about it all over the place. And the devs haven't done shit about it. which is why I joked about beating them with baseball bats until they do fix it. Because apparently, extreme disgust with the horrible performance characteristic of their creation doesn't seem to motivate them one bit. As a contrast, I can install the Google Desktop indexer (on the dual core system), and I never notice it is there. It indexes roughly the same scope of data (I think). It never runs so that I am aware it's indexing. My other apps carry on with no noticeable impact on performance. I see a few people here saying Beagle runs fine for them with no noticeable impact on performance... how? I've struggled with Beagle since it first appeared on the openSUSE scene. I have seen it's appalling impact on performance over several installs on several different hardware configurations. Not once have I seen it work in any measure that could be considered good. Personally, I think they're either lying, or not paying attention. I will continue to try it out with each new install I do, but... i don't hold out a lot of hope. I've kind of lumped it in with zmd... another app that is on my search and destroy list for a new install. Once those two apps are gone from a default install the computer works great with openSUSE. Yep! Why these system-resource hogs which offer functionality which is .. peripheral at best... are installed by default is utterly insane. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Rajko M. wrote: On Monday 14 January 2008 02:22:18 am Clayton wrote: I see a few people here saying Beagle runs fine for them with no noticeable impact on performance... how? It seems that you monitor Beagle in a first time after installation. Though there is pops up note telling that computer will be slower in a first few minutes. Later on you shouldn't notice indexing. Though that it should be optional as you suggested in another post as the version is 'beagle-0.2.18-30' which by any interparetation of version string is early development. On the other hand, how many people will ever attempt to test software with so low version (except Linux users)? Which brings back the question... Why in the hell is Beagle part of the default installation? it's 0.2 level, and it runs like complete crap. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 The Monday 2008-01-14 at 12:35 +0100, Clayton wrote: I let Beagle run longer than 24h on the dual core system. System response remained horrible. A friend installed 10.2 and then updated everything including Beagle... it ran for a couple of weeks with Beagle killing his system performance before he called and asked what was wrong. So, I am not talking 30 seconds of annoyance here... this is days of uptime on fast machines... and weeks on slower machines. And that is a bug you may report. There was a link with instructions. with basically no data, but about 1.2TB of data on other mount points. Which you could tell beagle not to index. That's a lot of data. True, but a significant portion of it is video. Mostly very large files eating up a lot of that diskspace not millions of small text files that need to be indexed. Indexing 2 or 3 hundred binary video files should not take that long. It indexes content, not just filenames, but I don't think it does anything with multimedia files. - -- Cheers, Carlos E. R. -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHi1wctTMYHG2NR9URAn2RAJ0TvSAGDPJPqgCzVdoJbPFjEAFplgCdG7eJ 0RHxXbXaiIUUzZeozbwVnPo= =OXiA -END PGP SIGNATURE- -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Monday 14 January 2008 05:03:11 am Aaron Kulkis wrote: Rajko M. wrote: On Monday 14 January 2008 02:22:18 am Clayton wrote: I see a few people here saying Beagle runs fine for them with no noticeable impact on performance... how? It seems that you monitor Beagle in a first time after installation. Though there is pops up note telling that computer will be slower in a first few minutes. Later on you shouldn't notice indexing. Though that it should be optional as you suggested in another post as the version is 'beagle-0.2.18-30' which by any interpretation of version string is early development. On the other hand, how many people will ever attempt to test software with so low version (except Linux users)? Which brings back the question... Why in the hell is Beagle part of the default installation? it's 0.2 level, and it runs like complete crap. The problem is that I don't see any bug report mentioned here, and right now I can't help. -- Regards, Rajko -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Monday 14 January 2008 05:35:30 am Clayton wrote: I see a few people here saying Beagle runs fine for them with no noticeable impact on performance... how? It seems that you monitor Beagle in a first time after installation. Though there is pops up note telling that computer will be slower in a first few minutes. Later on you shouldn't notice indexing. I let Beagle run longer than 24h on the dual core system. System response remained horrible. A friend installed 10.2 and then updated everything including Beagle... it ran for a couple of weeks with Beagle killing his system performance before he called and asked what was wrong. So, I am not talking 30 seconds of annoyance here... this is days of uptime on fast machines... and weeks on slower machines. version is 'beagle-0.2.18-30' which by any interpretation of version string is early development. On the other hand, how many people will ever attempt to test software with so low version (except Linux users)? Test by choice is a good thing... lots of us here install from Factory just to see what works. I have a VM I do that in all the time. Lots of things break and I have to roll back to a previous snapshot (which is why I like to use a VM instead of a native system) Setting it as part of the default install makes the new users test it as well. That isn't giving the new user a lot of choice. with basically no data, but about 1.2TB of data on other mount points. Which you could tell beagle not to index. That's a lot of data. True, but a significant portion of it is video. Mostly very large files eating up a lot of that diskspace not millions of small text files that need to be indexed. Indexing 2 or 3 hundred binary video files should not take that long. Clayton, here is the link to couple of bugs for 10.3 with word 'beagle' in description please look at it. Add your own if none does not describe your experience. https://bugzilla.novell.com/buglist.cgi?query_format=advancedshort_desc_type=allwordssubstrshort_desc=beaglelong_desc_type=fulltextlong_desc=classification=openSUSEproduct=openSUSE+10.3bug_file_loc_type=allwordssubstrbug_file_loc=status_whiteboard_type=allwordssubstrstatus_whiteboard=keywords_type=anywordskeywords=bug_status=NEWbug_status=ASSIGNEDbug_status=NEEDINFObug_status=REOPENEDemailassigned_to1=1emailtype1=substringemail1=emailassigned_to2=1emailreporter2=1emailqa_contact2=1emailcc2=1emailtype2=substringemail2=bugidtype=includebug_id=votes=chfieldfrom=chfieldto=Nowchfieldvalue=cmdtype=doitorder=Reuse+same+sort+as+last+timefield0-0-0=nooptype0-0-0=noopvalue0-0-0= Endless complains here have no much sense. It makes complains look like a lobbing, that is trying to discredit Beagel in order to advance some other technology. I have few machines and I can't confirm 99% CPU usage 100% of time on any of them. On this one it is comming up every few seconds (5-10s) and uses 10-30% of CPU. I have no time right now to retreive numbers for other machines, but it is just not that intrusive to make machine run slow. I use no special tweaking, it is just stock 10.3 installation. -- Regards, Rajko -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Yes, BTW, beagle is eating my CPU too... and I really would like to see this feature _disabled by default_ in the upcoming openSUSE 11.0. -- -Alexey Eremenko Technologov -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Mon, 2008-01-14 at 22:30 +0200, Alexey Eremenko wrote: Yes, BTW, beagle is eating my CPU too... and I really would like to see this feature _disabled by default_ in the upcoming openSUSE 11.0. -- -Alexey Eremenko Technologov As has been said before, that is something that a BUG needs to be filed. If you are not sure how to do this, if you'll provide us with somemore info, we could help you with that. Disabling a major feature of the desktop should NOT be an option on the table. -- Kevin Yo Dupuy | Public Email: [EMAIL PROTECTED] Happy New Year from Yo.media! -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Mon, 2008-01-14 at 06:03 -0500, Aaron Kulkis wrote: Rajko M. wrote: On Monday 14 January 2008 02:22:18 am Clayton wrote: I see a few people here saying Beagle runs fine for them with no noticeable impact on performance... how? It seems that you monitor Beagle in a first time after installation. Though there is pops up note telling that computer will be slower in a first few minutes. Later on you shouldn't notice indexing. Though that it should be optional as you suggested in another post as the version is 'beagle-0.2.18-30' which by any interparetation of version string is early development. On the other hand, how many people will ever attempt to test software with so low version (except Linux users)? Which brings back the question... Why in the hell is Beagle part of the default installation? it's 0.2 level, and it runs like complete crap. As we know, in the open source community, version numbers can mean different things to different projects. Beagle is NOT a project meant for testing, it is a project for use in user's desktops. -- Kevin Yo Dupuy | Public Email: [EMAIL PROTECTED] Happy New Year from Yo.media! -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Mon, 2008-01-14 at 12:35 +0100, Clayton wrote: I see a few people here saying Beagle runs fine for them with no noticeable impact on performance... how? It seems that you monitor Beagle in a first time after installation. Though there is pops up note telling that computer will be slower in a first few minutes. Later on you shouldn't notice indexing. I let Beagle run longer than 24h on the dual core system. System response remained horrible. A friend installed 10.2 and then updated everything including Beagle... it ran for a couple of weeks with Beagle killing his system performance before he called and asked what was wrong. So, I am not talking 30 seconds of annoyance here... this is days of uptime on fast machines... and weeks on slower machines. version is 'beagle-0.2.18-30' which by any interparetation of version string is early development. On the other hand, how many people will ever attempt to test software with so low version (except Linux users)? Test by choice is a good thing... lots of us here install from Factory just to see what works. I have a VM I do that in all the time. Lots of things break and I have to roll back to a previous snapshot (which is why I like to use a VM instead of a native system) Setting it as part of the default install makes the new users test it as well. That isn't giving the new user a lot of choice. with basically no data, but about 1.2TB of data on other mount points. Which you could tell beagle not to index. That's a lot of data. True, but a significant portion of it is video. Mostly very large files eating up a lot of that diskspace not millions of small text files that need to be indexed. Indexing 2 or 3 hundred binary video files should not take that long. C. Since noone seems to have brought this up before... http://beagle-project.org/Troubleshooting_CPU -- Kevin Yo Dupuy | Public Email: [EMAIL PROTECTED] Happy New Year from Yo.media! -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Mon, 2008-01-14 at 03:57 -0500, Aaron Kulkis wrote: Clayton wrote: Anyone else seeing Beagle really kill performance? I have disabled it and my machine finally is perky, but every now and then, I find it in memory again. How do I arange it to chew up less memory and CPU or kill it once and for all? Usually this indicates you have a problematic file (usually its broken or corrupt) that causes the index helper to go into a loop while indexing. See http://beagle-project.org/Troubleshooting_CPU for instructions on how to report such a bug. A bit late to the discussion here... I also have to kill Beagle every time I do an install. I tried it again with the 10.3 install I did this weekend. It sucked up so much of my system resources that I could barely do anything else... this is on a *clean* default install (not an upgrade) on an AMD Athlon 64 X2 3800+ with 2Gb of RAM, a /home with basically no data, but about 1.2TB of data on other mount points. My CPU.. both cores.. were running about 99%. RAM was full, and swap was filling up as well. The whole computer was grinding to a halt. When I finally managed top open a terminal and run top... Beagle was there consuming 100% of everything it could. I left Beagle run for a while... an afternoon... and it never changed. Kept my CPU nice and toasty warm though. In the end I sopped the daemon, and removed every trace of Beagle I could find. The result... the computer is back to normal. The 10.3 install is noticeably faster than the previous 10.2 install (also without Beagle) and I'm happy.. .although a bit puzzled how it is that anyone finds Beagle useable. This is typical behavior, and the developers have known about it for a long time, because there are written complaints about it all over the place. And the devs haven't done shit about it. which is why I joked about beating them with baseball bats until they do fix it. Because apparently, extreme disgust with the horrible performance characteristic of their creation doesn't seem to motivate them one bit. As a contrast, I can install the Google Desktop indexer (on the dual core system), and I never notice it is there. It indexes roughly the same scope of data (I think). It never runs so that I am aware it's indexing. My other apps carry on with no noticeable impact on performance. I see a few people here saying Beagle runs fine for them with no noticeable impact on performance... how? I've struggled with Beagle since it first appeared on the openSUSE scene. I have seen it's appalling impact on performance over several installs on several different hardware configurations. Not once have I seen it work in any measure that could be considered good. Personally, I think they're either lying, or not paying attention. I will continue to try it out with each new install I do, but... i don't hold out a lot of hope. I've kind of lumped it in with zmd... another app that is on my search and destroy list for a new install. Once those two apps are gone from a default install the computer works great with openSUSE. Yep! Why these system-resource hogs which offer functionality which is .. peripheral at best... are installed by default is utterly insane. Aaron... I would really like to thank you for calling me a liar and ignorant. To say the devs are lazy and not doing anything is pure stupidity on your part. I'm sorry, you need to go to the Beagle list and talk to them if this is your problem. Desktop Search is NOT a peripheral tool... for many users who have started to use it, it is just as important to the system use as a file manager. Just ask users of Beagle for a long time, or longtime users of Mac OS X Tiger who have used Spotlight. If you do not use it, then you are more than welcome to uninstall it. You, I believe, are NOT welcome to remove the application from users, particularly new users who like it and may not know how to install it manually. -- Kevin Yo Dupuy | Public Email: [EMAIL PROTECTED] Happy New Year from Yo.media! -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Kevin Dupuy wrote: On Mon, 2008-01-14 at 22:30 +0200, Alexey Eremenko wrote: Yes, BTW, beagle is eating my CPU too... and I really would like to see this feature _disabled by default_ in the upcoming openSUSE 11.0. -- -Alexey Eremenko Technologov As has been said before, that is something that a BUG needs to be filed. If you are not sure how to do this, if you'll provide us with somemore info, we could help you with that. Disabling a major feature of the desktop should NOT be an option on the table. If an automaker had a device which caused their cars to use 10x as much gas, and lose 95%+ power, nobody would call it a feature. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Joe Sloan wrote: Aaron Kulkis wrote: Sloan wrote: Gary Baribault wrote: And just for your general information, with Beagle installed, SuSE was using 1.5Gig of Memory and 256Meg of Swap. After removing Beagle and a reboot, I have 887Meg free and no swap used, I would think that Beagle qualifies as a HOG. I don't care what it offers as an advantage, it isn't worth that, (I have about 2 Gigs of EMail it was indexing) Well, I've never rebooted after nuking beagle, but I've always noticed similar benefits - Unfortunately beagle got my attention because whilst playing quake 3 arena online I would experience annoying pauses during game play, lasting up to half a second. Naturally I would often be fragged during these comatose periods. Removing beagle and zmd brings back silky smooth performance. It would be great if the beagle devs could take a page from the boinc playbook, and only use CPU when it is not being used by other apps. That's done by the scheduler in the kernel, not the app. And it's not just CPU usage... beagle stresses the entire system (especially clogging up filesystem I/O with a flood of outstanding requests which keep the disk-heads moving all the time, and introducing a big delay in disk I/O responsiveness for any other process (both already running and new processes). If you look at boinc, it's possible to do this with clever programming, completely in userland. When we've run boinc, the load average on the box rises to the point that you'd think the box is in trouble, judging from load average indications, but then you notice that everything is still quite responsive. clever programming is something I was a big fan of when I was in high school and college in the 1980's. Then I got out into the real world, where clever code usually translates very quickly into unmaintainable code. I used to LOVE programming in assembly language. And it would be very good if the world was static and unchanging. HOWEVER...the real world is constantly changing, and code written in assembly langauge has a very VERY short lifespan -- making it suitable only for embedded systems...and only then on short production runs for items which you WILL NOT be producing the product 2 years later. (or if you're doing something with really REALLY unusual constraints, like trying to fit a program and it's data into the on-board memory of a Motorola 68HC11 so that you have full use of ALL of the I/O ports as such, without having to use any as address/data busses. I haven't looked at the boinc code, but it's not new (I ran such tests in 2002 or so), and it's cross platform code, nothing linux specific there and all userspace, no kernel code. Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Saturday 22 December 2007 09:49, Aaron Kulkis wrote: Joe Sloan wrote: ... If you look at boinc, it's possible to do this with clever programming, completely in userland. When we've run boinc, the load average on the box rises to the point that you'd think the box is in trouble, judging from load average indications, but then you notice that everything is still quite responsive. clever programming is something I was a big fan of when I was in high school and college in the 1980's. Then I got out into the real world, where clever code usually translates very quickly into unmaintainable code. You're talking about gratuitous cleverness. And you're right to say that its rarely justifiable, but there's also sophistication, which can be justified, though it _must_ be justified and justified by something other than the author's or designer's sense of self-satisfaction. So using sufficiently sophisticated techniques to regulate the demand on system resources produced by, say the Beagle indexer (or a BOINC client task or the Google Desktop indexer) is justified, because for personal desktops or workstations, interference with the user's tasks is unacceptable. But it's also true that there's tremendous amounts of unused cycles and I/O bandwidth. So a program like the Beagle indexer has to be smart enough about the load it offers to get its work done as quickly as possible without interfering with interactive use. I'm pretty sure this is what Joe meant by clever programming in this context. ... Randall Schulz -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Sunday 13 January 2008 11:07, Randall R Schulz wrote: So a program like the Beagle indexer has to be smart enough about the load it offers to get its work done as quickly as possible without interfering with interactive use. By the way, I'm not saying I think Beagle does (or doesn't) need to do more or better along these lines. I don't have enough experience or information to have such an opinion. ... Randall Schulz -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Kevin Dupuy wrote: On Mon, 2007-12-17 at 10:53 -0500, Aaron Kulkis wrote: Gary Baribault wrote: Hi all, Anyone else seeing Beagle really kill performance? I have disabled it and my machine finally is perky, but every now and then, I find it in memory again. How do I arange it to chew up less memory and CPU or kill it once and for all? You can't. It's one of the stupidest pieces of software that I've ever seen in my entire 27 years of computing. Even putting it at the lowest priority nice value doesn't help, because the damn thing floods the disk drive with a constant stream of disk-head seeks, which interferes with any other process's attempts to use the disk drive by making it basically, stand in line with all of beagle's worthless disk I/O requests. The maintainers for beagle should be taken out and kneecapped, or beat about the head with a police baton. Every day Until they remove every trace of it from human existence, except for snippets preserved for teaching purposes, under the topic of DO NOT DO THESE THINGS in a background services programs I'm starting to get tired of the hating on Beagle's devs. Excuse me...has anyone really expressed HATRED towards the Beagle dev... Or have we just been noting that the software suffers from a VERY SERIOUS DEFICIENCY -- see Constructive Criticism hating on should be banished from the English language, because it is so widely abused by people who can't handle the fact that not everything or everybody is worthy of pure, unqualified praise now and forever. Most of the time when it is used, as in the posted example, it's just a sign of trying to suppress valid criticism. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Beagle performance: end-user problem? (was Re: [opensuse] Beagle under 10.3 is really eating up my CPU)
Linda Walsh wrote: The Saturday 2007-12-22 at 16:06 +0100, Anders Johansson wrote: The real solution here is to find and fix the bug that causes beagle to allocate so much memory. It doesn't happen on all systems. ---. Without question, this is the best solution. Anders Johansson wrote: I wouldn't say probably. It shouldn't be par for the course for an application to not check return values from memory allocation functions Aaron Kulkis wrote: As I said earlier... the whole thing is poorly written. --- And you say this based on??? It's performance. It's SUPPOSED to be an unobtrusive background process, but can cripple a high-end machine through extreme resource-hogging behavior. Unfriendly behavior is the very definition of poor code. Do you have _any_ expertise in the source? Not at all, but that's not even the issue. We're not talking about there's a faster way to sort this data sorts of tweaks...we're talking about code which is widely known to effectively cripple any system it runs on, due to nothing more than the files it's handling, and how many there are. (if you do, sorry, but your attitude is not productive). It _appears_ you know nothing about how it is written but are basing your opinion on its behavior in certain configurations. I'm basing my evaluation on how it performs, and its impact on system performancewhich is the ultimate standard of judging any code as to whether it is to be judged good enough or not. This is why I made a comment about *not* using swap on a system -- if you are using swap on any regular basis (my threshhold is using swap, *anytime*, during 'normal' day-to-day usage), you are running applications that are too big for for your machine. Now whether this is due to poor application memory usage OR is due to poor-planning on the part of the system owner depends, in large part, on promises or expectations set by the developer or owner of the program. I've got 2 GB of DDR2 on this machine, and 3 GB of swap, on a Centrino Core Duo running at 800 MHz with two internal 100 GB SATA drives, of which 140 GB is used by my SuSE installation. And a running beagle process makes this machine absolutely unusableeven without a GUI. Certainly, if I am running on a system with 128M of memory with a 650MHz mobile CPU, and load a full suse 10.x desktop with full features, I am asking that I be shot in the foot. If a release (like suse10.2, or windows 98) says it will run best with 1GB-mem + a 1GHz processor and my machine has 2GB+a 2GHz processor and the release runs like a dog -- then I'd say it is the fault of the release packager (they made the choice of what packages to include 'by default'). Certainly if the *end user* chooses to run more applications than their computer can comfortably fit in memory, how can the application developer account for this. Beagle will grow and grow and grow until it uses all available swap. It appears that the only way to satisfy beagle's appetite appears to be to have enough memory to load up all of your home directory tree into it. But I don't know of any motherboard sold for under US $10,000 that can accomodate 60 GB of ram. Beagle should be scrapped and started over from the ground up, starting with the design assumption that it is to behave as an unobtrusive background process, not the current one which can take over the whole system with a feed me attitude as if the whole purpose for a computer and its data to even exist is to provide something for a beagle process to index. - Do you have documentation or direct knowledge of what the design goals were? If not, how do you know it wasn't designed that way? I'm saying that the design goals either were not met, or they were utterly inappropriate. Something the beagle developers cannot know is how their application will be installed by release packagers. One example of an outstanding 'bug' (or feature depending on interpretation) that can affect beagle performance is how it is run by 'cron.daily'. From my own experience, under 9.3, the default is to run cron.daily 24 hours after it last ran -- but if something delays it running overnight (like the machine being off or suspended) it will run within 15 minutes of the machine becoming active. IT WON'T WAIT until the 'middle of the night', as you might want. This has nothing to do with beagle or its developers. I'm likely to be using this computer at all hours of the day and night... I wake up, get an idea, do something, and then go back to sleep... Ideally, the beagle indexing process would run once (either at night, or immediately if needed), and then be able to monitor the filesystems and directories for changes using the fam (famd) package. fam is another thing I've banished from my installations, for similar reasons. Maybe a good idea but it too suffers from poor implementation. The fam function (and as extended for directories and/or
Re: Capping memory (and, indirectly, swap) use of programs (was Re: [opensuse] Beagle under 10.3 is really eating up my CPU)
Linda Walsh wrote: Aaron Kulkis wrote: I just checked the man page for ulimit, and did a search for the string swap and it came up empty... so it looks like ulimit might not help herealthough ulimit -v can set an upper limit on virtual memory, --- You want to cap virtual memory usage, since virtual memory size is the total of a process's physical memory and swap-space use. As Carlos noted...that causes beagle to crash. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Sunday 23 December 2007 01:26:08 Carlos E. R. wrote: The Saturday 2007-12-22 at 18:09 +0100, Anders Johansson wrote: Actually, it helps solve it, sometimes. The application crashes, probably I wouldn't say probably. It shouldn't be par for the course for an application to not check return values from memory allocation functions That's not what I said. Look again: ] The application crashes, probably ] with an error, maybe a core dump or a backtrace, and it can be examined. There is a comma after the crashes. I didn't say that it probably crashes. I said that it will crash, and then probably will produce an error message. The idea is that an application that has a memory hole and uses a lot of memory doesn't crash, but by running with an ulimit it will crash when it hits the limit, ant then the error, coredump, or backtrace can be examined. Well, the idea might work, but I still say that you're too dogmatic. Even an application that forgets to free() memory can still test the return code from malloc() properly. It's not a given that it will crash. It might just exit normally, without a core being produced Anders -- Madness takes its toll -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Anders Johansson wrote: On Sunday 23 December 2007 01:26:08 Carlos E. R. wrote: The Saturday 2007-12-22 at 18:09 +0100, Anders Johansson wrote: Actually, it helps solve it, sometimes. The application crashes, probably I wouldn't say probably. It shouldn't be par for the course for an application to not check return values from memory allocation functions That's not what I said. Look again: ] The application crashes, probably ] with an error, maybe a core dump or a backtrace, and it can be examined. There is a comma after the crashes. I didn't say that it probably crashes. I said that it will crash, and then probably will produce an error message. The idea is that an application that has a memory hole and uses a lot of memory doesn't crash, but by running with an ulimit it will crash when it hits the limit, ant then the error, coredump, or backtrace can be examined. Well, the idea might work, but I still say that you're too dogmatic. Even an application that forgets to free() memory can still test the return code from malloc() properly. It's not a given that it will crash. It might just exit normally, without a core being produced] If it doesn't check return values from malloc() (or something which calls malloc()), it WILL crash due to a segmentation fault (trying to write to structure members through a structure-pointer which is NULL or -1. If it does check return it can still crash, but that would be due to other problems. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Carlos E. R. wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 The Thursday 2007-12-20 at 09:22 -1000, [EMAIL PROTECTED] wrote: Even if one nices and ionices beagle, there still is the matter of swap space getting filled. Usually it is when that happens when ugly thoughts against the developers spring to life. is there a swapnice or something for control of max swap space an app can use? If an application need memory, there is no memory free, and there is swap, the kernel will give it as much as it wants - unless you limit it with ulimit and friends. I just checked the man page for ulimit, and did a search for the string swap and it came up empty... so it looks like ulimit might not help herealthough ulimit -v can set an upper limit on virtual memory, and -d can cap the size of the process's data segment (although I have no idea of how that impacts malloc() or other dynamic memory allocation methods. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
[EMAIL PROTECTED] wrote: On Wednesday 19 December 2007 01:02:56 pm Aaron Kulkis wrote: ah... ionice seems to be a new invention. I'll give it a try when I build a new system this month. Even if one nices and ionices beagle, there still is the matter of swap space getting filled. Usually it is when that happens when ugly thoughts against the developers spring to life. is there a swapnice or something for control of max swap space an app can use? not on my 10.1 system. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Friday 21 December 2007 01:06:38 [EMAIL PROTECTED] wrote: On Thursday 20 December 2007 09:38:46 am Carlos E. R. wrote: The Thursday 2007-12-20 at 09:22 -1000, [EMAIL PROTECTED] wrote: Even if one nices and ionices beagle, there still is the matter of swap space getting filled. Usually it is when that happens when ugly thoughts against the developers spring to life. is there a swapnice or something for control of max swap space an app can use? If an application need memory, there is no memory free, and there is swap, the kernel will give it as much as it wants - unless you limit it with ulimit and friends. -- Cheers, Carlos E. R. a quick info ulimit reveals that perhaps the option RLIMIT_RSS can tell beagle to not fill all ram and swap, no? if so, then a tiny little script could totally control every aspect of resourse utilization by beagle. has anyone tried that? even i could do a small script with nice, ionice and ulimit only... The real solution here is to find and fix the bug that causes beagle to allocate so much memory. It doesn't happen on all systems. When it happens, a bug should be opened, so that the developers can investigate it. It shouldn't be hidden away by tricks like that, it should be exposed so it can be solved once and for all Anders -- Madness takes its toll -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 The Saturday 2007-12-22 at 16:06 +0100, Anders Johansson wrote: [ulimit] The real solution here is to find and fix the bug that causes beagle to allocate so much memory. It doesn't happen on all systems. When it happens, a bug should be opened, so that the developers can investigate it. It shouldn't be hidden away by tricks like that, it should be exposed so it can be solved once and for all Actually, it helps solve it, sometimes. The application crashes, probably with an error, maybe a core dump or a backtrace, and it can be examined. If it doesn't crash, there is nothing to report. - -- Cheers, Carlos E. R. -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHbUK+tTMYHG2NR9URAglVAJ0ZmRSqUDp9M1Fx2NlfUhVud7qP9QCfciB6 w0c77ulPFsQUYWoHdJhKay8= =h6Ok -END PGP SIGNATURE- -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Saturday 22 December 2007 18:00:43 Carlos E. R. wrote: The Saturday 2007-12-22 at 16:06 +0100, Anders Johansson wrote: [ulimit] The real solution here is to find and fix the bug that causes beagle to allocate so much memory. It doesn't happen on all systems. When it happens, a bug should be opened, so that the developers can investigate it. It shouldn't be hidden away by tricks like that, it should be exposed so it can be solved once and for all Actually, it helps solve it, sometimes. The application crashes, probably I wouldn't say probably. It shouldn't be par for the course for an application to not check return values from memory allocation functions Anders -- Madness takes its toll -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Anders Johansson wrote: On Saturday 22 December 2007 18:00:43 Carlos E. R. wrote: The Saturday 2007-12-22 at 16:06 +0100, Anders Johansson wrote: [ulimit] The real solution here is to find and fix the bug that causes beagle to allocate so much memory. It doesn't happen on all systems. When it happens, a bug should be opened, so that the developers can investigate it. It shouldn't be hidden away by tricks like that, it should be exposed so it can be solved once and for all Actually, it helps solve it, sometimes. The application crashes, probably I wouldn't say probably. It shouldn't be par for the course for an application to not check return values from memory allocation functions As I said earlier... the whole thing is poorly written. Beagle should be scrapped and started over from the ground up, starting with the design assumption that it is to behave as an unobtrusive background process, not the current one which can take over the whole system with a feed me attitude as if the whole purpose for a computer and its data to even exist is to provide something for a beagle process to index. Anders -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Sloan wrote: Aaron Kulkis wrote: Joe Sloan wrote: Examples, please? What would be the security advantage of typing /opt/SunWzztop/bin/top every time, instead of top, with /opt/SunWzztop/bin in the path? [EMAIL PROTECTED]:~ which top /usr/bin/top [EMAIL PROTECTED]:~ Why isn't top in /usr/bin where it belongs? Our solaris boxes always had top and similar utilities in /opt. Maybe it's different on solaris 10 (haven't checked), but historically, all the old school unix vendors shipped pretty bare bones systems from a usability point of view, and any useful extras tended to go either in /opt or /usr/local. We observe a few rules of thumb for the PATH of root - . is verboten, and world writable directories are out (actually any directory writable by someone other than root is suspect). Other than those restrictions, we prefer to use the path variable to make life less awkward and tedious. And since top is trustworthy, just put a symbolic link in /usr/bin to the top executable. Admittedly, it's a patch up, but it's superior to playing around with the $PATH variable(*) all the time (*) for one, it doesn't help much for anyone who has a currently running login with open shell windows, unless they log off, or at the very least, shut all the shell windows, and open up new ones. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Ken Schneider wrote: Aaron Kulkis pecked at the keyboard and wrote: Joe Sloan wrote: root's UID. Examples, please? What would be the security advantage of typing /opt/SunWzztop/bin/top every time, instead of top, with /opt/SunWzztop/bin in the path? [EMAIL PROTECTED]:~ which top /usr/bin/top [EMAIL PROTECTED]:~ Why isn't top in /usr/bin where it belongs? Looks like it is to me. I should have asked, Joe, why isn't YOUR system's 'top' command in /usr/bin where it belongs? I'm asking why on his machines it's in some bizarro, non-standard location. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Joe Sloan wrote: Dave Howorth wrote: Joe Sloan wrote: in fact anything in /opt or /usr/local usually wasn't even in root's path I should hope not! root should have a strictly limited path to limit security exploits. I'm not sure the lack of finish characteristic of the old school unices provides the security benefits you perceive. In any event the question becomes, is it lack of polish, or a carefully calculated attempt to limit the damage done by a bad superuser? The fact that not just roots path, but everybodys path seems a bit anemic on the old school unices, strongly hints at the former. No, it's the latter. IMHO the general user's path should certainly include commonly used utilities located in non-world-writable directories. You have to understand, Joe, that anything in /opt is non-standard, BY DEFINITION. IF it was standard, then it would be in /usr or /usr/bin or /usr/X11/bin or something similar. top is an edge case... some Unix vendors include it in their software distributions (HP, for instance), and some don't. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Aaron Kulkis wrote: And since top is trustworthy, just put a symbolic link in /usr/bin to the top executable. Admittedly, it's a patch up, but it's superior to playing around with the $PATH variable(*) all the time (*) for one, it doesn't help much for anyone who has a currently running login with open shell windows, unless they log off, or at the very least, shut all the shell windows, and open up new ones. I wasn't going to bring this up since it's more than anyone ever wanted to know about PATH, but we actually do put /usr/local/bin and /opt/bin in the path once, then make symbolic links for all the appropriate /opt binaries into /opt/bin, or /usr/local binaries into /usr/local/bin, for the reasons that you mentioned. SuSE packages tend to have these nice little scripts that drop into /etc/profile.d/, so we don't have to think much about that sort of thing anymore. Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
suse10.[23] beagle indexing automatically uses ionice if installed (was Re: [opensuse] Beagle under 10.3 is really eating up my CPU)
Aaron Kulkis wrote: ah... ionice seems to be a new invention. I'll give it a try when I build a new system this month. --- I just checked suse10.2 packages... If you have the ionice program (/usr/bin/ionice) installed under SuSE-10.3 OR -10.2, daily beagle indexing will automatically use it to set the beagle-indexing process to lowest (idle) priority using the command format ionice -c 3. The relevant lines in file /etc/cron.daily/beagle-crawl-system (fr suse10.3, beagle-0.2.12-28): lines 65-70: 65 IONICE=`which ionice 2/dev/null` 66 if [ -n $IONICE ]; then 67 IONICE=$IONICE -c 3 68 fi 69 70 eval nice -n 19 $IONICE su -s /bin/bash $CRAWL_USER -c \MONO_SH ARED_DIR=$MONO_SHARED_DIR /usr/sbin/beagle-build-index --target /var/cache/b eagle/indexes/$CRAWL_INDEX_NAME $OPTIONS $CRAWL_PATHS\ /dev/null 21 71 fi (don't know about 10.1, SLED, or 10.0 as I don't have those packages locally accessible, but shouldn't be too hard for someone to verify)... Linda -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: suse10.[23] beagle indexing automatically uses ionice if installed (was Re: [opensuse] Beagle under 10.3 is really eating up my CPU)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 * Linda Walsh [EMAIL PROTECTED] [12-22-07 18:18]: I just checked suse10.2 packages... If you have the ionice program (/usr/bin/ionice) installed under SuSE-10.3 OR -10.2, daily beagle indexing will automatically use it to set the beagle-indexing process to lowest (idle) priority using the command format ionice -c 3. The relevant lines in file /etc/cron.daily/beagle-crawl-system (fr suse10.3, beagle-0.2.12-28): lines 65-70: 65 IONICE=`which ionice 2/dev/null` 66 if [ -n $IONICE ]; then 67 IONICE=$IONICE -c 3 68 fi 69 70 eval nice -n 19 $IONICE su -s /bin/bash $CRAWL_USER -c \MONO_SH ARED_DIR=$MONO_SHARED_DIR /usr/sbin/beagle-build-index --target /var/cache/b eagle/indexes/$CRAWL_INDEX_NAME $OPTIONS $CRAWL_PATHS\ /dev/null 21 71 fi (don't know about 10.1, SLED, or 10.0 as I don't have those packages locally accessible, but shouldn't be too hard for someone to verify)... 10.1 is the same - -- Patrick Shanahan Plainfield, Indiana, USAHOG # US1244711 http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 Registered Linux User #207535@ http://counter.li.org -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.2 (GNU/Linux) iD8DBQFHbZ98ClSjbQz1U5oRAm/eAKCDNriXzNmnCaMGJUWGSPLLEagOhgCfb0/l iX7RAIZ0r6hvxWroWeI330o= =8caX -END PGP SIGNATURE- -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Beagle performance: end-user problem? (was Re: [opensuse] Beagle under 10.3 is really eating up my CPU)
The Saturday 2007-12-22 at 16:06 +0100, Anders Johansson wrote: The real solution here is to find and fix the bug that causes beagle to allocate so much memory. It doesn't happen on all systems. ---. Without question, this is the best solution. Anders Johansson wrote: I wouldn't say probably. It shouldn't be par for the course for an application to not check return values from memory allocation functions Aaron Kulkis wrote: As I said earlier... the whole thing is poorly written. --- And you say this based on??? Do you have _any_ expertise in the source? (if you do, sorry, but your attitude is not productive). It _appears_ you know nothing about how it is written but are basing your opinion on its behavior in certain configurations. This is why I made a comment about *not* using swap on a system -- if you are using swap on any regular basis (my threshhold is using swap, *anytime*, during 'normal' day-to-day usage), you are running applications that are too big for for your machine. Now whether this is due to poor application memory usage OR is due to poor-planning on the part of the system owner depends, in large part, on promises or expectations set by the developer or owner of the program. Certainly, if I am running on a system with 128M of memory with a 650MHz mobile CPU, and load a full suse 10.x desktop with full features, I am asking that I be shot in the foot. If a release (like suse10.2, or windows 98) says it will run best with 1GB-mem + a 1GHz processor and my machine has 2GB+a 2GHz processor and the release runs like a dog -- then I'd say it is the fault of the release packager (they made the choice of what packages to include 'by default'). Certainly if the *end user* chooses to run more applications than their computer can comfortably fit in memory, how can the application developer account for this. Beagle should be scrapped and started over from the ground up, starting with the design assumption that it is to behave as an unobtrusive background process, not the current one which can take over the whole system with a feed me attitude as if the whole purpose for a computer and its data to even exist is to provide something for a beagle process to index. - Do you have documentation or direct knowledge of what the design goals were? If not, how do you know it wasn't designed that way? Something the beagle developers cannot know is how their application will be installed by release packagers. One example of an outstanding 'bug' (or feature depending on interpretation) that can affect beagle performance is how it is run by 'cron.daily'. From my own experience, under 9.3, the default is to run cron.daily 24 hours after it last ran -- but if something delays it running overnight (like the machine being off or suspended) it will run within 15 minutes of the machine becoming active. IT WON'T WAIT until the 'middle of the night', as you might want. This has nothing to do with beagle or its developers. Ideally, the beagle indexing process would run once (either at night, or immediately if needed), and then be able to monitor the filesystems and directories for changes using the fam (famd) package. The fam function (and as extended for directories and/or devices) monitors when any change is done to its monitored file-system objects, then calls listening programs to process the new or changed objects as they are changed on disk. Ideally, you would then need no 'batch' updating, but such would be done in bits throughout the day as monitored files are changed. That being said, if a system doesn't have the OS support or resources needed to run 'famd' without without degradation, the system will still be painful to use (shorthand: be unusable). Be careful about global generalization about a product being bad, though, though, just because it doesn't run well in a particular situation. Linda -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Capping memory (and, indirectly, swap) use of programs (was Re: [opensuse] Beagle under 10.3 is really eating up my CPU)
Aaron Kulkis wrote: I just checked the man page for ulimit, and did a search for the string swap and it came up empty... so it looks like ulimit might not help herealthough ulimit -v can set an upper limit on virtual memory, --- You want to cap virtual memory usage, since virtual memory size is the total of a process's physical memory and swap-space use. L -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 The Saturday 2007-12-22 at 18:09 +0100, Anders Johansson wrote: Actually, it helps solve it, sometimes. The application crashes, probably I wouldn't say probably. It shouldn't be par for the course for an application to not check return values from memory allocation functions That's not what I said. Look again: ] The application crashes, probably ] with an error, maybe a core dump or a backtrace, and it can be examined. There is a comma after the crashes. I didn't say that it probably crashes. I said that it will crash, and then probably will produce an error message. The idea is that an application that has a memory hole and uses a lot of memory doesn't crash, but by running with an ulimit it will crash when it hits the limit, ant then the error, coredump, or backtrace can be examined. - -- Cheers, Carlos E. R. -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHbasrtTMYHG2NR9URArIgAKCUI9NyG9hRhD25ZuonZQ/BHlMALgCgkXWu eJY+OmRrUbhlqSr1oM7NPgw= =ND5g -END PGP SIGNATURE- -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Beagle performance: end-user problem? (was Re: [opensuse] Beagle under 10.3 is really eating up my CPU)
On Saturday 22 December 2007 07:29:04 pm Linda Walsh wrote: The Saturday 2007-12-22 at 16:06 +0100, Anders Johansson wrote: The real solution here is to find and fix the bug that causes beagle to allocate so much memory. It doesn't happen on all systems. ---. Without question, this is the best solution. Anders Johansson wrote: I wouldn't say probably. It shouldn't be par for the course for an application to not check return values from memory allocation functions Aaron Kulkis wrote: As I said earlier... the whole thing is poorly written. --- And you say this based on??? Do you have _any_ expertise in the source? (if you do, sorry, but your attitude is not productive). It _appears_ you know nothing about how it is written but are basing your opinion on its behavior in certain configurations. This is why I made a comment about *not* using swap on a system -- if you are using swap on any regular basis (my threshhold is using swap, *anytime*, during 'normal' day-to-day usage), you are running applications that are too big for for your machine. Now whether this is due to poor application memory usage OR is due to poor-planning on the part of the system owner depends, in large part, on promises or expectations set by the developer or owner of the program. Certainly, if I am running on a system with 128M of memory with a 650MHz mobile CPU, and load a full suse 10.x desktop with full features, I am asking that I be shot in the foot. If a release (like suse10.2, or windows 98) says it will run best with 1GB-mem + a 1GHz processor and my machine has 2GB+a 2GHz processor and the release runs like a dog -- then I'd say it is the fault of the release packager (they made the choice of what packages to include 'by default'). Certainly if the *end user* chooses to run more applications than their computer can comfortably fit in memory, how can the application developer account for this. Beagle should be scrapped and started over from the ground up, starting with the design assumption that it is to behave as an unobtrusive background process, not the current one which can take over the whole system with a feed me attitude as if the whole purpose for a computer and its data to even exist is to provide something for a beagle process to index. - Do you have documentation or direct knowledge of what the design goals were? If not, how do you know it wasn't designed that way? Something the beagle developers cannot know is how their application will be installed by release packagers. One example of an outstanding 'bug' (or feature depending on interpretation) that can affect beagle performance is how it is run by 'cron.daily'. From my own experience, under 9.3, the default is to run cron.daily 24 hours after it last ran -- but if something delays it running overnight (like the machine being off or suspended) it will run within 15 minutes of the machine becoming active. IT WON'T WAIT until the 'middle of the night', as you might want. This has nothing to do with beagle or its developers. Ideally, the beagle indexing process would run once (either at night, or immediately if needed), and then be able to monitor the filesystems and directories for changes using the fam (famd) package. The fam function (and as extended for directories and/or devices) monitors when any change is done to its monitored file-system objects, then calls listening programs to process the new or changed objects as they are changed on disk. Ideally, you would then need no 'batch' updating, but such would be done in bits throughout the day as monitored files are changed. That being said, if a system doesn't have the OS support or resources needed to run 'famd' without without degradation, the system will still be painful to use (shorthand: be unusable). Be careful about global generalization about a product being bad, though, though, just because it doesn't run well in a particular situation. Linda thanks, i think that about covers a broad spectrum of possible perceived problems... Howard -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: Beagle performance: end-user problem? (was Re: [opensuse] Beagle under 10.3 is really eating up my CPU)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 The Saturday 2007-12-22 at 16:29 -0800, Linda Walsh wrote: .. This is why I made a comment about *not* using swap on a system -- if you are using swap on any regular basis (my threshhold is using swap, *anytime*, during 'normal' day-to-day usage), you are running applications that are too big for for your machine. Tsk, tsk... I have a machine with 32 MiB of memory and about 1 GiB swap, and there were an application that filled more than half of it; and it run :-P The system is 7.3 and the app was yast (you, actually). It had a big memory hole (known bug). Without that much swap the update would simply crash, and new memory was no longer available for the machine. Still, it worked. A statement such as any swap usage is bad is not always correct for every body and every circumstance. It will not be as fast as having more ram, but... it works. Swap was designed for such a use. If designers thought that swap is a bad thing (R), they would not have designed kernel 2.6 with swap enabled. They would remove the swapping code and tell us to buy more ram instead. Hardware makers would be very happy. - -- Cheers, Carlos E. R. -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHbciPtTMYHG2NR9URAsnEAJ0cOhIJd5oyjDBBCglqUtwGK8tJXgCfSOK1 Zi76fauIDD77N9loNYpiQAc= =R0o3 -END PGP SIGNATURE- -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Beagle should be scrapped and started over from the ground up, starting with the design assumption that it is to behave as an unobtrusive background process, not the current one which can take over the whole system with a feed me attitude as if the whole purpose for a computer and its data to even exist is to provide something for a beagle process to index. This is scary and difficult to admit, but aaron is right on this one. mind you, JUST THIS ONE... punk d. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Joe Sloan wrote: Aaron Kulkis wrote: Joe Sloan wrote: I had to edit the paths even more extensively in hpux, solaris or aix - in fact anything in /opt or /usr/local usually wasn't even in root's path AND THEY'RE NOT SUPPOSED TO BE!!! That's a security risk. Root is for ADMINISTRATIVE use, not running apps. I'm not sure what you're getting at - are you saying it's a security risk for root to run top on solaris or hpux? A shame really, one of my favorite apps. How about linux? It's in root's path on every linux distro I've seen. But that is getting completely away from the point, which was that suse provides for root a fully functional path, but removes e.g. /sbin and /usr/sbin from the path of non-root users, which needs fixing up. - but I expect to spend time fixing things up to make those OSes Hat to say it, but no, you were possibly introducing security holes into those systems. Very few apps in /opt or /usr/local are ever tested for safety under root's UID. Examples, please? What would be the security advantage of typing /opt/SunWzztop/bin/top every time, instead of top, with /opt/SunWzztop/bin in the path? [EMAIL PROTECTED]:~ which top /usr/bin/top [EMAIL PROTECTED]:~ Why isn't top in /usr/bin where it belongs? -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Aaron Kulkis wrote: Joe Sloan wrote: Examples, please? What would be the security advantage of typing /opt/SunWzztop/bin/top every time, instead of top, with /opt/SunWzztop/bin in the path? [EMAIL PROTECTED]:~ which top /usr/bin/top [EMAIL PROTECTED]:~ Why isn't top in /usr/bin where it belongs? Our solaris boxes always had top and similar utilities in /opt. Maybe it's different on solaris 10 (haven't checked), but historically, all the old school unix vendors shipped pretty bare bones systems from a usability point of view, and any useful extras tended to go either in /opt or /usr/local. We observe a few rules of thumb for the PATH of root - . is verboten, and world writable directories are out (actually any directory writable by someone other than root is suspect). Other than those restrictions, we prefer to use the path variable to make life less awkward and tedious. Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Aaron Kulkis pecked at the keyboard and wrote: Joe Sloan wrote: root's UID. Examples, please? What would be the security advantage of typing /opt/SunWzztop/bin/top every time, instead of top, with /opt/SunWzztop/bin in the path? [EMAIL PROTECTED]:~ which top /usr/bin/top [EMAIL PROTECTED]:~ Why isn't top in /usr/bin where it belongs? Looks like it is to me. -- Ken Schneider SuSe since Version 5.2, June 1998 -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Wednesday 19 December 2007 01:02:56 pm Aaron Kulkis wrote: ah... ionice seems to be a new invention. I'll give it a try when I build a new system this month. Even if one nices and ionices beagle, there still is the matter of swap space getting filled. Usually it is when that happens when ugly thoughts against the developers spring to life. is there a swapnice or something for control of max swap space an app can use? -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 The Thursday 2007-12-20 at 09:22 -1000, [EMAIL PROTECTED] wrote: Even if one nices and ionices beagle, there still is the matter of swap space getting filled. Usually it is when that happens when ugly thoughts against the developers spring to life. is there a swapnice or something for control of max swap space an app can use? If an application need memory, there is no memory free, and there is swap, the kernel will give it as much as it wants - unless you limit it with ulimit and friends. - -- Cheers, Carlos E. R. -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHasTGtTMYHG2NR9URAv1mAJsHH7KNX/h2wedv4x+BfZjtzJpXtwCaAoCc JvbHXDt000Qx0VIwpFaH15g= =Ss3n -END PGP SIGNATURE- -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: [opensuse] Beagle under 10.3 is really eating up my CPU
It wouldn't be too hard to modify ionice to allow non-super users to decrease their priorities, and tell them to naff off if they try to increase priorities, then all it would need to be is suid root. Or you could write a wrapper to do the same. -Original Message- From: Randall R Schulz [mailto:[EMAIL PROTECTED] Sent: Thursday, 20 December 2007 2:36 p.m. To: opensuse@opensuse.org Subject: Re: [opensuse] Beagle under 10.3 is really eating up my CPU On Wednesday 19 December 2007 16:42, Carlos E. R. wrote: The Wednesday 2007-12-19 at 16:29 -0800, Randall R Schulz wrote: Unfortunately, that's right. You need to be root to increase and even decrease I/O priority using program ionice. How odd. Thankfully, I'm Lord and Master of my machines! So am I, but I don't want to run apps as root. Actually, that's not necessary. You can always become root for the purpose of exercising privileged operations and then go back to being a regular user to carry out everyday tasks. E.g.: su root -c 'ionice -c 3 su nonRootUser -c ultimate command' Or, if you're already root: ionice -c 3 su nonRootUser -c ultimate command -- Cheers, Carlos E. R. Randall Schulz -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
RE: [opensuse] Beagle under 10.3 is really eating up my CPU
lets say /opt/SunWzztop/bin is not first on your path, and for some reason a directory in your path was writable by some malicious user - if they put a top in there, then all of a sudden when you type in top, expecting to get /opt/SunWzztop/bin/top, you get another top instead. And this top does funky stuff like rm -rf / whoops. -Original Message- From: Joe Sloan [mailto:[EMAIL PROTECTED] Sent: Thursday, 20 December 2007 5:30 p.m. To: opensuse@opensuse.org Subject: Re: [opensuse] Beagle under 10.3 is really eating up my CPU --snip-- - but I expect to spend time fixing things up to make those OSes Hat to say it, but no, you were possibly introducing security holes into those systems. Very few apps in /opt or /usr/local are ever tested for safety under root's UID. Examples, please? What would be the security advantage of typing /opt/SunWzztop/bin/top every time, instead of top, with /opt/SunWzztop/bin in the path? Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Philip Dowie wrote: lets say /opt/SunWzztop/bin is not first on your path, and for some reason a directory in your path was writable by some malicious user - if they put a top in there, then all of a sudden when you type in top, expecting to get /opt/SunWzztop/bin/top, you get another top instead. And this top does funky stuff like rm -rf / whoops. There are certain conventions which must be observed. For instance . in the path is a non-no, for obvious reasons, and world-writeable directories should not be in the path either. Nobody here is suggesting otherwise, least of all me. My original point still stands, that it is useful for /sbin and /usr/sbin to be in the path of normal users. Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Thursday 20 December 2007 09:38:46 am Carlos E. R. wrote: The Thursday 2007-12-20 at 09:22 -1000, [EMAIL PROTECTED] wrote: Even if one nices and ionices beagle, there still is the matter of swap space getting filled. Usually it is when that happens when ugly thoughts against the developers spring to life. is there a swapnice or something for control of max swap space an app can use? If an application need memory, there is no memory free, and there is swap, the kernel will give it as much as it wants - unless you limit it with ulimit and friends. -- Cheers, Carlos E. R. a quick info ulimit reveals that perhaps the option RLIMIT_RSS can tell beagle to not fill all ram and swap, no? if so, then a tiny little script could totally control every aspect of resourse utilization by beagle. has anyone tried that? even i could do a small script with nice, ionice and ulimit only... -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Wednesday 19 December 2007 04:54:01 am Linda Walsh wrote: Rajko M. wrote: ... --- Might try making sure the cfq block algorithm is being used, then set 'beagle' to run at lowest priority (nice -19 beagle-start-script). info:nice Nicenesses range at least from -20 (resulting in the most favorable scheduling) through 19 (the least favorable). === Right -- and to set priority 19, you use nice -19 prog, but you knew that, right? :-) I knew what info tells: nice -n niceness program In practice it accepts also nice -19 program and nice --19 program for negative values. It is just the way you quoted command line that confused me. It appeared to me as text. Beagle is actually silent helper in background. --- No such thing in standard linux. The cpu nice doesn't affect the disk-io priority unless you have the non-standard cfq scheduling algorithm enabled. The default when I installed 10.2 recently, I believe was the 'anticipatory' deadline. Unfortunately, while it may be good for server workloads, and better for throughput, 'cfq' is better for interactive use. A background process can easily saturate the disk if it runs at full speed (even if process is 'niced' down). In addition to using 'cfq' ('fair' queuing), you can run the beagle processes in 'batch' priority -- which will be below normal user processes. Beagled -- is that a background indexer and beagle-helper is to aid foreground searches? Or...why are there two processes? You question is the answer :-) http://beagle-project.org/Main_Page I can't see Joe Shaw, jumping in discussion. He knows the best about the beagle. There was problem in initial release of 10.2 where it was started Beagle and mandb. Running both on same hard disk made system sluggish, and that happened on every boot, few minutes after GUI was up. --- Mandb finishes after a few minutes - virtually never runs -- can't see how it would drag down beagle... It was 2 processes competing about single hard disk and few minutes computer was in very bad shape, and it was computer that had no problems with quite a few applications running at the same time. I experienced the problem and initially I was against Beagle, but Joe helped to find real reason for problems and ever since I can't see it as a resource hog. I'm just happy user. Course if it was a real beagle, just wave some treats in front of it -- it'll get active feisty! :-) Sure :-) Many users noticed Beagle and missed to see mandb, and since removing one of programs that were competing for hard disk access made situation much better, Beagle earned bad reputation and it is still comming back trough Google search. Beagle-helper runs now with nice=19, so it is already the lowest priority and it will not make problem even on initial indexing. The beagled runs with nice=7, so it is also below most processes in the system. --- If people think it is a problem, why not run it at 19? But a disk-bound nice-19 process can still hog the system. Yes, but this is not the case with Beagle here. The problem was solved in 10.2, and should not appear in 10.3, unless Mono jumped in as regression factor. -- Regards, Rajko -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Rajko M. wrote: On Sunday 16 December 2007 04:19:48 pm Linda Walsh wrote: Gary Baribault wrote: Hi all, Anyone else seeing Beagle really kill performance? I have disabled it and my machine finally is perky, but every now and then, I find it in memory again. How do I arange it to chew up less memory and CPU or kill it once and for all? --- Might try making sure the cfq block algorithm is being used, then set 'beagle' to run at lowest priority (nice -19 beagle-start-script). info:nice Nicenesses range at least from -20 (resulting in the most favorable scheduling) through 19 (the least favorable). === Right -- and to set priority 19, you use nice -19 prog, but you knew that, right? :-) That should help it not use so much CPU, and, if cfq is working well, it should set beagle's disk priority to near lowest as well. Of course, if beagle is using 500MB and you only have 512MB, you are likely to get alot of swapping. I'd also wonder, does beagle use alot of resources during some initial full-index phase, after which it can run with less resources as it does incremental updates...? BTW -- anyone compared it to swish (another full-system indexing util with web-based interface). Linda Beagle is actually silent helper in background. --- No such thing in standard linux. The cpu nice doesn't affect the disk-io priority unless you have the non-standard cfq scheduling algorithm enabled. The default when I installed 10.2 recently, I believe was the 'anticipatory' deadline. Unfortunately, while it may be good for server workloads, and better for throughput, 'cfq' is better for interactive use. A background process can easily saturate the disk if it runs at full speed (even if process is 'niced' down). In addition to using 'cfq' ('fair' queuing), you can run the beagle processes in 'batch' priority -- which will be below normal user processes. Beagled -- is that a background indexer and beagle-helper is to aid foreground searches? Or...why are there two processes? There was problem in initial release of 10.2 where it was started Beagle and mandb. Running both on same hard disk made system sluggish, and that happened on every boot, few minutes after GUI was up. --- Mandb finishes after a few minutes - virtually never runs -- can't see how it would drag down beagle... Course if it was a real beagle, just wave some treats in front of it -- it'll get active feisty! :-) Many users noticed Beagle and missed to see mandb, and since removing one of programs that were competing for hard disk access made situation much better, Beagle earned bad reputation and it is still comming back trough Google search. Beagle-helper runs now with nice=19, so it is already the lowest priority and it will not make problem even on initial indexing. The beagled runs with nice=7, so it is also below most processes in the system. --- If people think it is a problem, why not run it at 19? But a disk-bound nice-19 process can still hog the system. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Linda Walsh wrote: No such thing in standard linux. The cpu nice doesn't affect the disk-io priority unless you have the non-standard cfq scheduling algorithm enabled. The default when I installed 10.2 recently, I believe was the 'anticipatory' deadline. Unfortunately, while it may be good for server workloads, and better for throughput, 'cfq' is better for interactive use. A background process can easily saturate the disk if it runs at full speed (even if process is 'niced' down). I'm curious and didn't find an answer with a quick google. How does one inspect what algorithm is in use and/or change it? Is it a runtime option or build-time? Thanks, Dave -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
[EMAIL PROTECTED] wrote: On Sunday 16 December 2007 05:56:05 pm Joe Sloan wrote: David C. Rankin wrote: rpm -e $(rpm -qa | grep beagle) works nicely It looks good, but it won't remove beagle because kerry needs it. But in general I agree with your elegant approach. Joe You can safely remove kerry and anything beagle. and make some noise about it, perhaps it can attract the attention of developers that develop bloated software. In my mind it is really sad that anything not related to gaming or heavy duty engineering simulations abuses hardware thousands of times harder than it could or should... it used to be that open source software was a lean and mean fighting machine, now the typical linucs install is about 2-3x that of an xp partition, don't know anything about vista. and running the proggies often brings up situations like beagle or a software update, much better than 10.2 but still awful timewise, on dual core or even quad core cpus with oodles of ram! d. I believe that some of the developers who are converts from LoseDOS have brought their perverted ways with them :-( such as their belief that even in an environment capable of multi-processing, users really only do one thing at a time. And look at how many people tolerate the 50%+ CPU usage of Norton's system monitor on their Lose-DOS systems. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Joe Sloan wrote: Peter Van Lone wrote: On Dec 16, 2007 4:39 PM, Joe Sloan [EMAIL PROTECTED] wrote: The standard procedure for me on any new suse build is to nuke beagle completely Beagle doesn't give me any problems at all ... and I use it often to find stuff ... I guess if I was using the box as a server I would remove it, but otherwise it seems to me that removing it automatically is more a statement concerning how little you value desktop search, more than a statement about Beagle itself? (along with fixing the broken non-root paths, and installing What do you consider the broken non-root paths? Just curious. I have to fix the path to deal with the complaints of users who complain that e.g. ifconfig isn't installed, or who maybe have to type a full path for common commands. There's IMHO no reason a normal desktop user shouldn't be able to run many commands which reside in /sbin or /usr/sbin. Unix has always been like that, too. It used to be that many admin commands that are now in /sbin or /usr/sbin used to be in /etc. And /etc was never in any user's path except for root's. Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Gary Baribault wrote: Hi all, Anyone else seeing Beagle really kill performance? I have disabled it and my machine finally is perky, but every now and then, I find it in memory again. How do I arange it to chew up less memory and CPU or kill it once and for all? You can't. It's one of the stupidest pieces of software that I've ever seen in my entire 27 years of computing. Even putting it at the lowest priority nice value doesn't help, because the damn thing floods the disk drive with a constant stream of disk-head seeks, which interferes with any other process's attempts to use the disk drive by making it basically, stand in line with all of beagle's worthless disk I/O requests. The maintainers for beagle should be taken out and kneecapped, or beat about the head with a police baton. Every day Until they remove every trace of it from human existence, except for snippets preserved for teaching purposes, under the topic of DO NOT DO THESE THINGS in a background services programs -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Joe Sloan wrote: Gary Baribault wrote: Hi all, Anyone else seeing Beagle really kill performance? I have disabled it and my machine finally is perky, but every now and then, I find it in memory again. How do I arange it to chew up less memory and CPU or kill it once and for all? The standard procedure for me on any new suse build is to nuke beagle completely (along with fixing the broken non-root paths, and installing the chronically missing rwhod) That means, specifically, doing an rpm -qa | grep beagle, nuking every resulting item and also any dependencies such as kerry or kio_beagle. yeah, those two helper programs are annoying as hell, too. Everyone associated with Beagle should be forced to program exclusively on Windows for the rest of their natural lives. Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Kevin Dupuy wrote: On Sun, 2007-12-16 at 21:46 -0600, Peter Van Lone wrote: On Dec 16, 2007 4:39 PM, Joe Sloan [EMAIL PROTECTED] wrote: The standard procedure for me on any new suse build is to nuke beagle completely Beagle doesn't give me any problems at all ... and I use it often to find stuff ... I guess if I was using the box as a server I would remove it, but otherwise it seems to me that removing it automatically is more a statement concerning how little you value desktop search, more than a statement about Beagle itself? (along with fixing the broken non-root paths, and installing What do you consider the broken non-root paths? Just curious. Peter I agree with you. I have no issue with Beagle, I think maybe the issues are related to the size of the Home folder...? I've never had a single system on which beagle wasn't an complete system hog... both CPU and flooding all of my disk drives with so many I/O requests that it made a system built on a set of 4 SCSI disks sound like a bunch of rabid chipmunks, and made the system have the same sort of responsiveness I endured when in college working on a 1 MB, 1 MHz VAX with 25 other people logged in at the same time, doing edit/compile/test-run cycles. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Aaron Kulkis wrote: Joe Sloan wrote: I have to fix the path to deal with the complaints of users who complain that e.g. ifconfig isn't installed, or who maybe have to type a full path for common commands. There's IMHO no reason a normal desktop user shouldn't be able to run many commands which reside in /sbin or /usr/sbin. Unix has always been like that, too. It used to be that many admin commands that are now in /sbin or /usr/sbin used to be in /etc. And /etc was never in any user's path except for root's. I had to edit the paths even more extensively in hpux, solaris or aix - in fact anything in /opt or /usr/local usually wasn't even in root's path - but I expect to spend time fixing things up to make those OSes work correctly. Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Joe Sloan wrote: in fact anything in /opt or /usr/local usually wasn't even in root's path I should hope not! root should have a strictly limited path to limit security exploits. Cheers, Dave -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Dave Howorth wrote: Joe Sloan wrote: in fact anything in /opt or /usr/local usually wasn't even in root's path I should hope not! root should have a strictly limited path to limit security exploits. I'm not sure the lack of finish characteristic of the old school unices provides the security benefits you perceive. In any event the question becomes, is it lack of polish, or a carefully calculated attempt to limit the damage done by a bad superuser? The fact that not just roots path, but everybodys path seems a bit anemic on the old school unices, strongly hints at the former. IMHO the general user's path should certainly include commonly used utilities located in non-world-writable directories. Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Gary Baribault wrote: And just for your general information, with Beagle installed, SuSE was using 1.5Gig of Memory and 256Meg of Swap. After removing Beagle and a reboot, I have 887Meg free and no swap used, I would think that Beagle qualifies as a HOG. I don't care what it offers as an advantage, it isn't worth that, (I have about 2 Gigs of EMail it was indexing) But but butis that *all the time*?... Or is it doing some initial indexing? I thought someone else said that it took a while (long time) to index their whole disk, but after it was done, it only needed to smaller, incremental overnight runs. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Anders Johansson wrote: Fortunately there is a way of doing it - with ionice you can set IO scheduling priority idle, and since it's the IO that kills you, it should be good enough --- I thought ionice only worked with the 'cfq' block scheduler. Is that the suse default? -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Aaron Kulkis wrote: Until they remove every trace of it from human existence, except for snippets preserved for teaching purposes, under the topic of DO NOT DO THESE THINGS in a background services programs --- Well it depends on your configuration, no doubt. I saw one user with 512M memory 1.5G swap. These days, an interactive desktop user should almost never have more swap than physical memory -- even having swap=memory is too slow in many cases if the swap is being used frequently ( once/week). Try setting the 'ionice' level of beagle to 'batch' (below interactive processes) and see if you have the problem. ? -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Wed, 2007-12-19 at 13:59 -0800, Linda Walsh wrote: These days, an interactive desktop user should almost never have more swap than physical memory -- even having swap=memory is too slow in many cases if the swap is being used frequently ( once/week). I have 2 GB of real memory and 10 GB of swap on my interactive desktop machine but I'm probably not a normal user :) When the swap gets used, it slows the machine down, but it's usually better than having the program crash. And the machine is still usable. Much worse (i.e. unusable) is when there's heavy file access, usually NFS, which is why I'm interested in the cfq scheduling! The symptoms then are almost as bad as when our one and only DNS server has a senior moment (but it should be getting a younger sibling soon). Cheers, Dave -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 The Monday 2007-12-17 at 20:24 +0100, Anders Johansson wrote: Fortunately there is a way of doing it - with ionice you can set IO scheduling priority idle, and since it's the IO that kills you, it should be good enough Only if you are root. - -- Cheers, Carlos E. R. -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHaaV4tTMYHG2NR9URAjquAJ0WUVdIlNyQPtFm7gQeeRIh9xGZbwCeM+q9 yUJtrXFoCB4PUd0pwIEF2FY= =5cre -END PGP SIGNATURE- -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Mon, 2007-12-17 at 10:53 -0500, Aaron Kulkis wrote: Gary Baribault wrote: Hi all, Anyone else seeing Beagle really kill performance? I have disabled it and my machine finally is perky, but every now and then, I find it in memory again. How do I arange it to chew up less memory and CPU or kill it once and for all? You can't. It's one of the stupidest pieces of software that I've ever seen in my entire 27 years of computing. Even putting it at the lowest priority nice value doesn't help, because the damn thing floods the disk drive with a constant stream of disk-head seeks, which interferes with any other process's attempts to use the disk drive by making it basically, stand in line with all of beagle's worthless disk I/O requests. The maintainers for beagle should be taken out and kneecapped, or beat about the head with a police baton. Every day Until they remove every trace of it from human existence, except for snippets preserved for teaching purposes, under the topic of DO NOT DO THESE THINGS in a background services programs I'm starting to get tired of the hating on Beagle's devs. I don't care if you don't like the software, although I do find it sad, since Beagle is actually very good, and I do know that it can cause some issues with some people, but don't hate on the actual people behind the software. If you hated Wal-Mart, do you go up to the greeter at the door and tell them their knees should be broken? Seriously, get back to the actual discussion about the software, and stop hating on the actual people behind the software. -- Kevin Yo Dupuy | Public Email: [EMAIL PROTECTED] Merry Christmas from Yo.media! -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Wednesday 19 December 2007 15:12, Carlos E. R. wrote: The Monday 2007-12-17 at 20:24 +0100, Anders Johansson wrote: Fortunately there is a way of doing it - with ionice you can set IO scheduling priority idle, and since it's the IO that kills you, it should be good enough Only if you are root. Is ionice different from (CPU) nice? I.e., is it not possible for any user to lower the priority (increase the absolute nice value) while only root can improve the priority (decrease the nice value)? -- Cheers, Carlos E. R. Randall Schulz -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 The Wednesday 2007-12-19 at 15:23 -0800, Randall R Schulz wrote: Only if you are root. Is ionice different from (CPU) nice? I.e., is it not possible for any user to lower the priority (increase the absolute nice value) while only root can improve the priority (decrease the nice value)? Unfortunately, that's right. You need to be root to increase and even decrease I/O priority using program ionice. - -- Cheers, Carlos E. R. -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHabFXtTMYHG2NR9URAo5oAJ0T4/+aNG8gSO13iW53NO8S4YhynQCaA1aZ zHLZnRR/3zoq7mDNQMBaMjw= =p6Bh -END PGP SIGNATURE- -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Sloan wrote: Gary Baribault wrote: And just for your general information, with Beagle installed, SuSE was using 1.5Gig of Memory and 256Meg of Swap. After removing Beagle and a reboot, I have 887Meg free and no swap used, I would think that Beagle qualifies as a HOG. I don't care what it offers as an advantage, it isn't worth that, (I have about 2 Gigs of EMail it was indexing) Well, I've never rebooted after nuking beagle, but I've always noticed similar benefits - Unfortunately beagle got my attention because whilst playing quake 3 arena online I would experience annoying pauses during game play, lasting up to half a second. Naturally I would often be fragged during these comatose periods. Removing beagle and zmd brings back silky smooth performance. It would be great if the beagle devs could take a page from the boinc playbook, and only use CPU when it is not being used by other apps. That's done by the scheduler in the kernel, not the app. And it's not just CPU usage... beagle stresses the entire system (especially clogging up filesystem I/O with a flood of outstanding requests which keep the disk-heads moving all the time, and introducing a big delay in disk I/O responsiveness for any other process (both already running and new processes). Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Wednesday 19 December 2007 15:47, Carlos E. R. wrote: The Wednesday 2007-12-19 at 15:23 -0800, Randall R Schulz wrote: Only if you are root. Is ionice different from (CPU) nice? I.e., is it not possible for any user to lower the priority (increase the absolute nice value) while only root can improve the priority (decrease the nice value)? Unfortunately, that's right. You need to be root to increase and even decrease I/O priority using program ionice. How odd. Thankfully, I'm Lord and Master of my machines! -- Cheers, Carlos E. R. RRS -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 The Wednesday 2007-12-19 at 16:29 -0800, Randall R Schulz wrote: Unfortunately, that's right. You need to be root to increase and even decrease I/O priority using program ionice. How odd. Thankfully, I'm Lord and Master of my machines! So am I, but I don't want to run apps as root. - -- Cheers, Carlos E. R. -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHabqQtTMYHG2NR9URAjWrAJ4tPSHtT3tKnamvjEcyKrhmC+hZFgCfR4SZ gyKowUHU65TdJde4NE2TvDI= =A3Y7 -END PGP SIGNATURE- -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Wednesday 19 December 2007 16:42, Carlos E. R. wrote: The Wednesday 2007-12-19 at 16:29 -0800, Randall R Schulz wrote: Unfortunately, that's right. You need to be root to increase and even decrease I/O priority using program ionice. How odd. Thankfully, I'm Lord and Master of my machines! So am I, but I don't want to run apps as root. Actually, that's not necessary. You can always become root for the purpose of exercising privileged operations and then go back to being a regular user to carry out everyday tasks. E.g.: su root -c 'ionice -c 3 su nonRootUser -c ultimate command' Or, if you're already root: ionice -c 3 su nonRootUser -c ultimate command -- Cheers, Carlos E. R. Randall Schulz -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 The Wednesday 2007-12-19 at 17:35 -0800, Randall R Schulz wrote: So am I, but I don't want to run apps as root. Actually, that's not necessary. You can always become root for the purpose of exercising privileged operations and then go back to being a regular user to carry out everyday tasks. E.g.: su root -c 'ionice -c 3 su nonRootUser -c ultimate command' Or, if you're already root: ionice -c 3 su nonRootUser -c ultimate command Ah! Ah! Didn't know that trick. I have to try that! [...] Doesn't work... nimrodel:~ # ionice -c 1 su - cer -c xine --verbose Who needs friends when you can sit alone in your room and drink? This is xine (X11 gui) - a free video player v0.99.5. (c) 2000-2007 The xine Team. Built with xine library 1.1.8 (1.1.8) Found xine library version: 1.1.8 (1.1.8). Cannot open display nimrodel:~ # Mayby doesn't like su -. Another try: nimrodel:~ # ionice -c 1 su cer -c xine --verbose Yes! Lets check on another terminal: nimrodel:~ # ps au | grep xine root 14954 0.0 0.1 3124 1208 pts/27 S+ 03:29 0:00 su cer -c xine --verbose cer 14955 2.8 2.7 205492 28944 pts/27 Sl+ 03:29 0:02 xine --verbose root 15122 0.0 0.0 2040 652 pts/20 S+ 03:30 0:00 grep xine nimrodel:~ # ionice -p 14955 realtime: prio 4 Good! It is runnin with realtime priority as user. I like that! Now I'll have to write this to a sudo rule, some how. - -- Cheers, Carlos E. R. -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.4-svn0 (GNU/Linux) iD8DBQFHadVptTMYHG2NR9URApzDAJ48R0AxaO2cevw1mSP53Tk8IvX9bwCdGJVx lUumOPNTtan8hgdyvsuQYKE= =6ce8 -END PGP SIGNATURE- -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Joe Sloan wrote: Aaron Kulkis wrote: Joe Sloan wrote: I have to fix the path to deal with the complaints of users who complain that e.g. ifconfig isn't installed, or who maybe have to type a full path for common commands. There's IMHO no reason a normal desktop user shouldn't be able to run many commands which reside in /sbin or /usr/sbin. Unix has always been like that, too. It used to be that many admin commands that are now in /sbin or /usr/sbin used to be in /etc. And /etc was never in any user's path except for root's. I had to edit the paths even more extensively in hpux, solaris or aix - in fact anything in /opt or /usr/local usually wasn't even in root's path AND THEY'RE NOT SUPPOSED TO BE!!! That's a security risk. Root is for ADMINISTRATIVE use, not running apps. - but I expect to spend time fixing things up to make those OSes Hat to say it, but no, you were possibly introducing security holes into those systems. Very few apps in /opt or /usr/local are ever tested for safety under root's UID. work correctly. Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Anders Johansson wrote: On Monday 17 December 2007 19:58:18 JP Rosevear wrote: Beagle already does nice itself and employ strategies for reducing work when the CPU is not idle. However I/O is a problem and non-root processes can't change their own I/O priority iirc. Actually, they can. All they need is CAP_SYS_ADMIN A bit silly to not allow a process to downgrade its own priority, with nice I All processes can downgrade their own priority. Nice with positive (lower priority) values is available to all processes. Only negative (higher priority) values need the root user ID. can set myself to absolute rock bottom. Even sillier is it to have the capability baked in with some real admin capabilities. It should be more fine grained But what beagle could do is start as root (or CAP_SYS_ADMIN), run ioprio_set(), then immediately drop root perms. Anders -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Linda Walsh wrote: Aaron Kulkis wrote: Until they remove every trace of it from human existence, except for snippets preserved for teaching purposes, under the topic of DO NOT DO THESE THINGS in a background services programs --- Well it depends on your configuration, no doubt. I saw one user with 512M memory 1.5G swap. These days, an interactive desktop user should almost never have more swap than physical memory -- even having swap=memory is too slow in many cases if the swap is being used frequently ( once/week). Try setting the 'ionice' level of beagle to 'batch' (below interactive processes) and see if you have the problem. ? ah... ionice seems to be a new invention. I'll give it a try when I build a new system this month. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Aaron Kulkis wrote: Sloan wrote: Gary Baribault wrote: And just for your general information, with Beagle installed, SuSE was using 1.5Gig of Memory and 256Meg of Swap. After removing Beagle and a reboot, I have 887Meg free and no swap used, I would think that Beagle qualifies as a HOG. I don't care what it offers as an advantage, it isn't worth that, (I have about 2 Gigs of EMail it was indexing) Well, I've never rebooted after nuking beagle, but I've always noticed similar benefits - Unfortunately beagle got my attention because whilst playing quake 3 arena online I would experience annoying pauses during game play, lasting up to half a second. Naturally I would often be fragged during these comatose periods. Removing beagle and zmd brings back silky smooth performance. It would be great if the beagle devs could take a page from the boinc playbook, and only use CPU when it is not being used by other apps. That's done by the scheduler in the kernel, not the app. And it's not just CPU usage... beagle stresses the entire system (especially clogging up filesystem I/O with a flood of outstanding requests which keep the disk-heads moving all the time, and introducing a big delay in disk I/O responsiveness for any other process (both already running and new processes). If you look at boinc, it's possible to do this with clever programming, completely in userland. When we've run boinc, the load average on the box rises to the point that you'd think the box is in trouble, judging from load average indications, but then you notice that everything is still quite responsive. I haven't looked at the boinc code, but it's not new (I ran such tests in 2002 or so), and it's cross platform code, nothing linux specific there and all userspace, no kernel code. Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Aaron Kulkis wrote: Joe Sloan wrote: I had to edit the paths even more extensively in hpux, solaris or aix - in fact anything in /opt or /usr/local usually wasn't even in root's path AND THEY'RE NOT SUPPOSED TO BE!!! That's a security risk. Root is for ADMINISTRATIVE use, not running apps. I'm not sure what you're getting at - are you saying it's a security risk for root to run top on solaris or hpux? A shame really, one of my favorite apps. How about linux? It's in root's path on every linux distro I've seen. But that is getting completely away from the point, which was that suse provides for root a fully functional path, but removes e.g. /sbin and /usr/sbin from the path of non-root users, which needs fixing up. - but I expect to spend time fixing things up to make those OSes Hat to say it, but no, you were possibly introducing security holes into those systems. Very few apps in /opt or /usr/local are ever tested for safety under root's UID. Examples, please? What would be the security advantage of typing /opt/SunWzztop/bin/top every time, instead of top, with /opt/SunWzztop/bin in the path? Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Wednesday 19 December 2007 23:51:42 Aaron Kulkis wrote: Anders Johansson wrote: On Monday 17 December 2007 19:58:18 JP Rosevear wrote: Beagle already does nice itself and employ strategies for reducing work when the CPU is not idle. However I/O is a problem and non-root processes can't change their own I/O priority iirc. Actually, they can. All they need is CAP_SYS_ADMIN A bit silly to not allow a process to downgrade its own priority, with nice I All processes can downgrade their own priority. Not I/O priority, no. Nice with positive (lower priority) values is available to all processes. Only negative (higher priority) values need the root user ID. Yes, that's what I said. Anders -- Madness takes its toll -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Well, it's not quite that easy... Beagle libraries are used by a bunch of installed programs, but I un-installed all of the rest and have my nice snappy thunderbird back .. :-) Thanks for all of the suggestions, made for fun reading Gary B Joe Sloan wrote: [EMAIL PROTECTED] wrote: On Sunday 16 December 2007 05:56:05 pm Joe Sloan wrote: David C. Rankin wrote: rpm -e $(rpm -qa | grep beagle) works nicely It looks good, but it won't remove beagle because kerry needs it. But in general I agree with your elegant approach. Joe You can safely remove kerry and anything beagle. Right, and I always remove kerry - I was just pointing out a flaw in the one-liner provided earlier as an example. and make some noise about it, perhaps it can attract the attention of developers that develop bloated software. In my mind it is really sad that anything not related to gaming or heavy duty engineering simulations abuses hardware thousands of times harder than it could or should... it used to be that open source software was a lean and mean fighting machine, now the typical linucs install is about 2-3x that of an xp partition, don't know anything about vista. and running the proggies often brings up situations like beagle or a software update, much better than 10.2 but still awful timewise, on dual core or even quad core cpus with oodles of ram! Well it still can be very lean and mean, but if you install suse, you have to do some work to get it that way. Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Sun, 2007-12-16 at 16:41 -0500, Gary Baribault wrote: Hi all, Anyone else seeing Beagle really kill performance? I have disabled it and my machine finally is perky, but every now and then, I find it in memory again. How do I arange it to chew up less memory and CPU or kill it once and for all? Usually this indicates you have a problematic file (usually its broken or corrupt) that causes the index helper to go into a loop while indexing. See http://beagle-project.org/Troubleshooting_CPU for instructions on how to report such a bug. -JP -- JP Rosevear [EMAIL PROTECTED] Novell, Inc. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Mon 17 December 07 07:39, Gary Baribault wrote: Well, it's not quite that easy... Beagle libraries are used by a bunch of installed programs, but I un-installed all of the rest and have my nice snappy thunderbird back .. :-) Thanks for all of the suggestions, made for fun reading Gary B Please don't top-post. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Kevin Dupuy wrote: Some people have a lot of trouble with Beagle, I don't. If you want to kill it off for good, you can uninstall it. What are your computer's specs, I'm trying to figure out why some are having issues and others aren't. I don't seem to have a problem with it. Oh. I'm thinking it is maybe the size of a user's /home. My condo is about 900 sq feet. ;-) -- Use OpenOffice.org http://www.openoffice.org -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
And just for your general information, with Beagle installed, SuSE was using 1.5Gig of Memory and 256Meg of Swap. After removing Beagle and a reboot, I have 887Meg free and no swap used, I would think that Beagle qualifies as a HOG. I don't care what it offers as an advantage, it isn't worth that, (I have about 2 Gigs of EMail it was indexing) Gary B Joe Sloan wrote: [EMAIL PROTECTED] wrote: On Sunday 16 December 2007 05:56:05 pm Joe Sloan wrote: David C. Rankin wrote: rpm -e $(rpm -qa | grep beagle) works nicely It looks good, but it won't remove beagle because kerry needs it. But in general I agree with your elegant approach. Joe You can safely remove kerry and anything beagle. Right, and I always remove kerry - I was just pointing out a flaw in the one-liner provided earlier as an example. and make some noise about it, perhaps it can attract the attention of developers that develop bloated software. In my mind it is really sad that anything not related to gaming or heavy duty engineering simulations abuses hardware thousands of times harder than it could or should... it used to be that open source software was a lean and mean fighting machine, now the typical linucs install is about 2-3x that of an xp partition, don't know anything about vista. and running the proggies often brings up situations like beagle or a software update, much better than 10.2 but still awful timewise, on dual core or even quad core cpus with oodles of ram! Well it still can be very lean and mean, but if you install suse, you have to do some work to get it that way. Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Mon, 2007-12-17 at 11:38 -0500, Gary Baribault wrote: And just for your general information, with Beagle installed, SuSE was using 1.5Gig of Memory and 256Meg of Swap. After removing Beagle and a reboot, I have 887Meg free and no swap used, I would think that Beagle qualifies as a HOG. I don't care what it offers as an advantage, it isn't worth that, (I have about 2 Gigs of EMail it was indexing) Well, without seeing the processes involved its hard to tell. A couple notes: a) corrupt files/attachments etc will often cause this (and it repeats every time it tries to index it again) b) the thunderbird backend was pretty inefficient until 0.3.1 (which is too new for 10.3) -JP -- JP Rosevear [EMAIL PROTECTED] Novell, Inc. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Gary Baribault wrote: And just for your general information, with Beagle installed, SuSE was using 1.5Gig of Memory and 256Meg of Swap. After removing Beagle and a reboot, I have 887Meg free and no swap used, I would think that Beagle qualifies as a HOG. I don't care what it offers as an advantage, it isn't worth that, (I have about 2 Gigs of EMail it was indexing) Well, I've never rebooted after nuking beagle, but I've always noticed similar benefits - Unfortunately beagle got my attention because whilst playing quake 3 arena online I would experience annoying pauses during game play, lasting up to half a second. Naturally I would often be fragged during these comatose periods. Removing beagle and zmd brings back silky smooth performance. It would be great if the beagle devs could take a page from the boinc playbook, and only use CPU when it is not being used by other apps. Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Monday 17 December 2007 12:22, Sloan wrote: It would be great if the beagle devs could take a page from the boinc playbook, and only use CPU when it is not being used by other apps. You mean like the windoze devs...? cpu timeslice should *never* be in the hands of app developers. The kernel schedules the cpu, and timeslice not app devs. (windoze never mind) However, a sysadmin can adjust the cpu timeslice for beagle, or any other cpu intensive app, so that they crawl along happily in the background. The apps under such control take longer to complete of course... but in the case of massive indexing like beagle --who cares? -- Kind regards, M Harris -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Mon, 2007-12-17 at 06:44 -0600, M Harris wrote: On Monday 17 December 2007 12:22, Sloan wrote: It would be great if the beagle devs could take a page from the boinc playbook, and only use CPU when it is not being used by other apps. You mean like the windoze devs...? cpu timeslice should *never* be in the hands of app developers. The kernel schedules the cpu, and timeslice not app devs. (windoze never mind) However, a sysadmin can adjust the cpu timeslice for beagle, or any other cpu intensive app, so that they crawl along happily in the background. The apps under such control take longer to complete of course... but in the case of massive indexing like beagle --who cares? Beagle already does nice itself and employ strategies for reducing work when the CPU is not idle. However I/O is a problem and non-root processes can't change their own I/O priority iirc. -JP -- JP Rosevear [EMAIL PROTECTED] Novell, Inc. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Monday 17 December 2007 13:44:20 M Harris wrote: On Monday 17 December 2007 12:22, Sloan wrote: It would be great if the beagle devs could take a page from the boinc playbook, and only use CPU when it is not being used by other apps. You mean like the windoze devs...? cpu timeslice should *never* be in the hands of app developers. The kernel schedules the cpu, and timeslice not app devs. (windoze never mind) What's windows got to do with it. If you have an application that requires lots of resources but has a very low priority, what's wrong with saying to the kernel only run this when nothing else is running? I think you're confusing nice with applications' getting priority when they're in the foreground (which is the windows strategy), but that's the other way around Fortunately there is a way of doing it - with ionice you can set IO scheduling priority idle, and since it's the IO that kills you, it should be good enough Anders -- Madness takes its toll -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
M Harris wrote: On Monday 17 December 2007 12:22, Sloan wrote: It would be great if the beagle devs could take a page from the boinc playbook, and only use CPU when it is not being used by other apps. You mean like the windoze devs...? Nope, I mean the boinc devs - windoze is not a worthy role model. cpu timeslice should *never* be in the hands of app developers. Who said anything about cpu timeslice? we're talking pure userland here. The kernel schedules the cpu, and timeslice not app devs. (windoze never mind) It's defintely in the hands of the app, in the sense that the app can *not* demand resources. However, a sysadmin can adjust the cpu timeslice for beagle, or any other cpu intensive app, so that they crawl along happily in the background. The apps under such control take longer to complete of course... but in the case of massive indexing like beagle --who cares? The problem is, Aunt Martha should not have to wrestle with that. There should be a sane default setting. Joe -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Monday 17 December 2007 19:58:18 JP Rosevear wrote: Beagle already does nice itself and employ strategies for reducing work when the CPU is not idle. However I/O is a problem and non-root processes can't change their own I/O priority iirc. Actually, they can. All they need is CAP_SYS_ADMIN A bit silly to not allow a process to downgrade its own priority, with nice I can set myself to absolute rock bottom. Even sillier is it to have the capability baked in with some real admin capabilities. It should be more fine grained But what beagle could do is start as root (or CAP_SYS_ADMIN), run ioprio_set(), then immediately drop root perms. Anders -- Madness takes its toll -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Monday 17 December 2007 13:24, Anders Johansson wrote: I think you're confusing nice with applications' getting priority when they're in the foreground (which is the windows strategy), but that's the other way around Yes... I was alluding to the windoze strategy, but not confused about nice... must poking fun at the windoze strategy... Fortunately there is a way of doing it - with ionice you can set IO scheduling priority idle, and since it's the IO that kills you, it should be good enough Yes... nice | ionice are our friends... -- Kind regards, M Harris -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Mon, 2007-12-17 at 09:18 -0500, JP Rosevear wrote: On Sun, 2007-12-16 at 16:41 -0500, Gary Baribault wrote: Hi all, Anyone else seeing Beagle really kill performance? I have disabled it and my machine finally is perky, but every now and then, I find it in memory again. How do I arange it to chew up less memory and CPU or kill it once and for all? Usually this indicates you have a problematic file (usually its broken or corrupt) that causes the index helper to go into a loop while indexing. See http://beagle-project.org/Troubleshooting_CPU for instructions on how to report such a bug. -JP -- JP Rosevear [EMAIL PROTECTED] Novell, Inc. I was looking for a page like that, thanks... -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Monday 17 December 2007 08:58:18 am JP Rosevear wrote: Beagle already does nice itself and employ strategies for reducing work when the CPU is not idle. However I/O is a problem and non-root processes can't change their own I/O priority iirc. -JP -- JP Rosevear [EMAIL PROTECTED] Novell, Inc. and what happens when it fills the swap space and brings *every* app to a grinding halt all while it nices itself to 19 or 38 or whatever? when things get as bad as beagle often makes them, the simplest thing to do is to throw the whole package away and start from scratch; yes, that would be hard, difficult and costly, but, a proper end product would more than pay for itself and would bring fame and fortune to the developer. some problems can not be solved simply by connecting dots among prefabricated libs. d. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Mon, 2007-12-17 at 20:32 +0100, Anders Johansson wrote: On Monday 17 December 2007 19:58:18 JP Rosevear wrote: Beagle already does nice itself and employ strategies for reducing work when the CPU is not idle. However I/O is a problem and non-root processes can't change their own I/O priority iirc. Actually, they can. All they need is CAP_SYS_ADMIN A bit silly to not allow a process to downgrade its own priority, with nice I can set myself to absolute rock bottom. Even sillier is it to have the capability baked in with some real admin capabilities. It should be more fine grained I concur. And in fact the system wide index thats shared for documentation, apps, etc does run as a privileged user and takes advantage of ionice (/etc/cron.daily/beagle-crawl-system). But what beagle could do is start as root (or CAP_SYS_ADMIN), run ioprio_set(), then immediately drop root perms. Have you met our security team? :-) All kidding aside, this sounds like an interesting idea, although indexing happens in a shorter lived indexing processes rather than a proper daemon so the indexer would have to do that itself. -JP -- JP Rosevear [EMAIL PROTECTED] Novell, Inc. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
On Sunday 16 December 2007 04:19:48 pm Linda Walsh wrote: Gary Baribault wrote: Hi all, Anyone else seeing Beagle really kill performance? I have disabled it and my machine finally is perky, but every now and then, I find it in memory again. How do I arange it to chew up less memory and CPU or kill it once and for all? --- Might try making sure the cfq block algorithm is being used, then set 'beagle' to run at lowest priority (nice -19 beagle-start-script). info:nice Nicenesses range at least from -20 (resulting in the most favorable scheduling) through 19 (the least favorable). That should help it not use so much CPU, and, if cfq is working well, it should set beagle's disk priority to near lowest as well. Of course, if beagle is using 500MB and you only have 512MB, you are likely to get alot of swapping. I'd also wonder, does beagle use alot of resources during some initial full-index phase, after which it can run with less resources as it does incremental updates...? BTW -- anyone compared it to swish (another full-system indexing util with web-based interface). Linda Beagle is actually silent helper in background. There was problem in initial release of 10.2 where it was started Beagle and mandb. Running both on same hard disk made system sluggish, and that happened on every boot, few minutes after GUI was up. Many users noticed Beagle and missed to see mandb, and since removing one of programs that were competing for hard disk access made situation much better, Beagle earned bad reputation and it is still comming back trough Google search. Beagle-helper runs now with nice=19, so it is already the lowest priority and it will not make problem even on initial indexing. The beagled runs with nice=7, so it is also below most processes in the system. -- Regards, Rajko -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Re: [opensuse] Beagle under 10.3 is really eating up my CPU
Beagle-helper runs now with nice=19, so it is already the lowest priority and it will not make problem even on initial indexing. The beagled runs with nice=7, so it is also below most processes in the system. -- Regards, Rajko it will still eat up the ram and fill the swap space. if that can be addressed as well, maybe there is some hope. d. -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
[opensuse] Beagle under 10.3 is really eating up my CPU
Hi all, Anyone else seeing Beagle really kill performance? I have disabled it and my machine finally is perky, but every now and then, I find it in memory again. How do I arange it to chew up less memory and CPU or kill it once and for all? Gary B -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]