Another question
Maybe I missed it, but what's the deal with the new kernels (2.4.0xxx) and shared memory? From top: CPU states: 5.2% user, 1.8% system, 0.0% nice, 93.0% idle Mem: 78592K av, 75848K used, 2744K free, 0K shrd, 1932K buff Swap: 185464K av, 11520K used, 173944K free 40308K cached Shared mem is always 0 ? Tim -- >< >> Tim Sailer (at home) >< Coastal Internet, Inc. << >> Network and Systems Operations >< PO Box 671 << >> http://www.buoy.com >< Ridge, NY 11961 << >> [EMAIL PROTECTED]/[EMAIL PROTECTED] >< (631) 476-3031 << ><
lvm
Has anyone had any luck getting the lvm stuff to work? Tim -- >< >> Tim Sailer (at home) >< Coastal Internet, Inc. << >> Network and Systems Operations >< PO Box 671 << >> http://www.buoy.com >< Ridge, NY 11961 << >> [EMAIL PROTECTED]/[EMAIL PROTECTED] >< (631) 476-3031 << ><
Re: AMD Duron CPU & Debian
On Tue, Aug 29, 2000 at 10:08:53PM -0600, Art Sackett wrote:
> ('compact' is sporting a 2.2.14pre-something
Dump typo from a poorly-skilled keyboard operator. 2.2.17pre-something.
--
Art Sackett
Re: AMD Duron CPU & Debian
On Thu, Aug 24, 2000 at 10:05:46PM +0100, Jonathan McDowell wrote:
>
> I had a machine with a 2940 lock up after a fair few SCSI bus resets. I
> compiled up the latest 2.2.17 pre release which has the latest driver in
> it and turned off tagged command queuing and haven't seen a problem
> since.
I finally bit the bullet a few days ago and installed potato on one of
the machines I was concerned about. Using the 'compact' images and a
network install, it worked well. It turns out that the 3C905C driver is
in the package and works well, and the AHA-2940-U2W just came right up
and ran without a problem. ('compact' is sporting a 2.2.14pre-something
kernel -- could be 2.2.16 would have barfed, I don't know.)
One thing it wouldn't do was reach out to the internet for the HTTP
installation, but it could have been something I did wrong, while in
a hurry. It would get to things on my local subnet, though, so I setup
a ProxyPass directive on Apache in another machine and sneak around
the problem. Once the installation was fully in, the new install had
no trouble reaching the internet. I don't know what I did wrong, aside
from being constantly interrupted and more-or-less autopiloting my way
through the procedure, trusting it to be a lot like it has been in the
past... it worked, I'm happy, can't ask for much more than that.
--
Art Sackett
Re: Another question
Check the Readme's with the Kernel source - there is actualy a device you have to mount in your fstab file (you know, for bootup;) that enables shared memory. It uses a "imaginary" mount point like /proc does. However, as far as I know, "free" and friends don't show the shared memory in use correctly (though df showes the mount point...) Sorry I can't remember the exact command, if you can't find it - please let me know and I will dig it up from a server I have it setup on. -Nathan > Maybe I missed it, but what's the deal with the new kernels (2.4.0xxx) > and shared memory? From top: > > CPU states: 5.2% user, 1.8% system, 0.0% nice, 93.0% idle > Mem: 78592K av, 75848K used, 2744K free, 0K shrd, 1932K buff > Swap: 185464K av, 11520K used, 173944K free 40308K cached > > > Shared mem is always 0 ? > > Tim > > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Another question
Maybe I missed it, but what's the deal with the new kernels (2.4.0xxx) and shared memory? From top: CPU states: 5.2% user, 1.8% system, 0.0% nice, 93.0% idle Mem: 78592K av, 75848K used, 2744K free, 0K shrd, 1932K buff Swap: 185464K av, 11520K used, 173944K free 40308K cached Shared mem is always 0 ? Tim -- >< >> Tim Sailer (at home) >< Coastal Internet, Inc. << >> Network and Systems Operations >< PO Box 671 << >> http://www.buoy.com >< Ridge, NY 11961 << >> [EMAIL PROTECTED][EMAIL PROTECTED] >< (631) 476-3031 << >< -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
lvm
Has anyone had any luck getting the lvm stuff to work? Tim -- >< >> Tim Sailer (at home) >< Coastal Internet, Inc. << >> Network and Systems Operations >< PO Box 671 << >> http://www.buoy.com >< Ridge, NY 11961 << >> [EMAIL PROTECTED][EMAIL PROTECTED] >< (631) 476-3031 << >< -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: AMD Duron CPU & Debian
On Tue, Aug 29, 2000 at 10:08:53PM -0600, Art Sackett wrote:
> ('compact' is sporting a 2.2.14pre-something
Dump typo from a poorly-skilled keyboard operator. 2.2.17pre-something.
--
Art Sackett
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: AMD Duron CPU & Debian
On Thu, Aug 24, 2000 at 10:05:46PM +0100, Jonathan McDowell wrote:
>
> I had a machine with a 2940 lock up after a fair few SCSI bus resets. I
> compiled up the latest 2.2.17 pre release which has the latest driver in
> it and turned off tagged command queuing and haven't seen a problem
> since.
I finally bit the bullet a few days ago and installed potato on one of
the machines I was concerned about. Using the 'compact' images and a
network install, it worked well. It turns out that the 3C905C driver is
in the package and works well, and the AHA-2940-U2W just came right up
and ran without a problem. ('compact' is sporting a 2.2.14pre-something
kernel -- could be 2.2.16 would have barfed, I don't know.)
One thing it wouldn't do was reach out to the internet for the HTTP
installation, but it could have been something I did wrong, while in
a hurry. It would get to things on my local subnet, though, so I setup
a ProxyPass directive on Apache in another machine and sneak around
the problem. Once the installation was fully in, the new install had
no trouble reaching the internet. I don't know what I did wrong, aside
from being constantly interrupted and more-or-less autopiloting my way
through the procedure, trusting it to be a lot like it has been in the
past... it worked, I'm happy, can't ask for much more than that.
--
Art Sackett
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: reiserfs & databases.
On Tue, Aug 29, 2000 at 04:36:23PM +0200, Dariush Pietrzak wrote: > but, there are some commercial databases which keep their data directly > on partitions ( this should be much better then any *fs including > reiserfs) and the weird part is that that direct-partition instalation > scheme seems to be a little bit slower that fs-based in benchmarks. > And this means that I'm missing something here, what is it that I haven't > thought about, anyone, any comments on this? If I understand your question, you're saying that RDBMs do benchmark faster using a native filesystems rather than rolling their own on a partition, and you're wondering why ... I would have to hazard a guess that the operating system disk cache and buffers are coming into play when you're using a native filesystem, but there's no caching when a "raw" partition is used. -- "Eschew Obfuscation" email:[EMAIL PROTECTED]http://incanus.net/~nnorman pgpQMv31j0vlY.pgp Description: PGP signature
Re: what is sufficient free memory?
Hello again, Robert Davies wrote: > Why wait? > > Run the command vmstat, and observe how much is paged in /out, what is the > scan rate? That indicates how hard the page stealer is looking for pages it > can free off. vmstat 10 is usually a goodish, number but if you can run it > a long time, longer sample times is useful. OK, here goes: procs memoryswap io system cpu r b w swpd free buff cache si sobibo incs us sy id 0 0 0 3560 1644 31188 8608 0 0 4 1 10417 0 0 99 > If you still have the 128MB in the machine, you could force Linux to ignore > it, using a boot parameter. Nah, the machine won't boot unless I remove it. > With squid, you probably want to lower the mem cache down to about 1/4 of > physical RAM, if it's higher than the default. Depending on usage perhaps > less than 16MB would be a waste of RAM, your call! presumably via /etc/squid.conf ? > If the suits don't blink, then perhaps you could investigate interleaving > swap over a number of disk partions, by using mkswap, and swap entries in > fstab all set with a priority=5, instead of the defaults which don't > interleave. We'll see, I'm still waiting for a larger barracuda. It'd be nice if I can have both in one shabbang! for now I'm fairly confident that with the adjustment of squid, my swap partition can manage until I get the 128MB. Or won't that be enough? I'll soon find out and tell everyone how it goes. Mabuhay! Erik
ppp configuration PROBLEM
Help: I am new to Debian and am unabel to get internet applications running after connecting with wvdial?? Perhaps someone can advise me where I can enter my ISP's DNS address? Any help will be gratly appeciated. Jim Clawson __ Do You Yahoo!? Yahoo! Mail - Free email you can access from anywhere! http://mail.yahoo.com/
Re: what is sufficient free memory?
> Thanks for all the reponses. From most of the replies, can I gather that > I'll have to observe my how much is being swapped to determine whether I > should immediately "up" the RAM back to 128MB? (and pester the tight-wad > suits who'll approve the requesition) Why wait? Run the command vmstat, and observe how much is paged in /out, what is the scan rate? That indicates how hard the page stealer is looking for pages it can free off. vmstat 10 is usually a goodish, number but if you can run it a long time, longer sample times is useful. If you still have the 128MB in the machine, you could force Linux to ignore it, using a boot parameter. With lilo, that would mean an append line with 'mem=64M' in it, or enter it at the lilo boot prompt. boot: linux mem=64M That way you can actually trial your system out, at the cost of a reboot, to see the affect. With squid, you probably want to lower the mem cache down to about 1/4 of physical RAM, if it's higher than the default. Depending on usage perhaps less than 16MB would be a waste of RAM, your call! If the suits don't blink, then perhaps you could investigate interleaving swap over a number of disk partions, by using mkswap, and swap entries in fstab all set with a priority=5, instead of the defaults which don't interleave. Rob
Re: what is sufficient free memory?
Hello All, Thanks for all the reponses. From most of the replies, can I gather that I'll have to observe my how much is being swapped to determine whether I should immediately "up" the RAM back to 128MB? (and pester the tight-wad suits who'll approve the requesition) To paint a better picture, here's an entire top screen: 70 processes: 69 sleeping, 1 running, 0 zombie, 0 stopped CPU states: 0.1% user, 0.7% system, 0.0% nice, 99.0% idle Mem: 63124K av, 60764K used, 2360K free, 38700K shrd, 32216K buff Swap: 104380K av, 3572K used, 100808K free 7772K cached PID USER PRI NI SIZE RSS SHARE STAT LIB %CPU %MEM TIME COMMAND 1047 beyonder 10 0 1032 1032 820 R 0 0.7 1.6 0:00 top 717 root 1 0 1340 1288 1040 S 0 0.1 2.0 0:01 sshd 1 root 0 0 460 460 388 S 0 0.0 0.7 0:06 init 2 root 0 0 00 0 SW 0 0.0 0.0 0:00 kflushd 3 root 0 0 00 0 SW 0 0.0 0.0 0:00 kupdate 4 root 0 0 00 0 SW 0 0.0 0.0 0:00 kpiod 5 root 0 0 00 0 SW 0 0.0 0.0 0:00 kswapd 6 root -20 -20 00 0 SW< 0 0.0 0.0 0:00 mdrecoveryd 297 root 1 0 480 472 388 S 0 0.0 0.7 0:00 syslogd 308 root 0 0 684 672 316 S 0 0.0 1.0 0:00 klogd 324 root 1 0 548 548 480 S 0 0.0 0.8 0:00 crond 340 root 15 0 472 468 392 S 0 0.0 0.7 0:00 inetd 356 root 0 0 1556 1244 596 S 0 0.0 1.9 0:00 named 366 root 0 0 772 700 604 S 0 0.0 1.1 0:01 sshd 394 root 0 0 516 508 412 S 0 0.0 0.8 0:00 automount 452 root 2 0 688 688 548 S 0 0.0 1.0 0:00 master 458 postfix0 0 812 812 648 S 0 0.0 1.2 0:00 qmgr 470 root 0 0 400 384 324 S 0 0.0 0.6 0:00 gpm 486 root 0 0 1368 1364 1288 S 0 0.0 2.1 0:00 httpd 490 nobody 0 0 1220 1172 1044 S 0 0.0 1.8 0:00 httpd 491 nobody 0 0 1224 1224 1096 S 0 0.0 1.9 0:00 httpd 492 nobody 0 0 964 964 872 S 0 0.0 1.5 0:00 httpd 493 nobody 0 0 964 964 872 S 0 0.0 1.5 0:00 httpd 494 nobody 0 0 1220 1220 1096 S 0 0.0 1.9 0:00 httpd 495 nobody 0 0 1268 1268 1120 S 0 0.0 2.0 0:00 httpd The stats on the swapping have been constant for the past hour or so, but notable too is that it's 4am here in the Philippines and there aren't any subscribers who've dialed in and engaged squid as they surf. I guess I'll be able to get a better idea later on when my subscribers start pouring in. Mabuhay kayong lahat at debian! (long live you all and debian) Erik
Re: reiserfs & databases.
On Tue, Aug 29, 2000 at 04:36:23PM +0200, Dariush Pietrzak wrote: > but, there are some commercial databases which keep their data directly > on partitions ( this should be much better then any *fs including > reiserfs) and the weird part is that that direct-partition instalation > scheme seems to be a little bit slower that fs-based in benchmarks. > And this means that I'm missing something here, what is it that I haven't > thought about, anyone, any comments on this? If I understand your question, you're saying that RDBMs do benchmark faster using a native filesystems rather than rolling their own on a partition, and you're wondering why ... I would have to hazard a guess that the operating system disk cache and buffers are coming into play when you're using a native filesystem, but there's no caching when a "raw" partition is used. -- "Eschew Obfuscation" email:[EMAIL PROTECTED]http://incanus.net/~nnorman PGP signature
Re: what is sufficient free memory?
On Tue, 29 Aug 2000, Nathan wrote: > Why not that cheap? Less than half of his ram is being used for actual > needed data. The rest is either free, in cache or in buffers which means > the memory isn't even close to being stressed on the system. The 3MB in that's exactly what i'm talking about. the rest is being used as cache, just because it's there, and once it's there and it's not needed for ``real'' data storage, it's being used as cache. nothing's cheap enough just to lay around unused ;) -- [-] ``And there are plenty of other innovative pieces of software such as Napster and ICQ.'' -- comment on ``Systems Software Research is Irrelevant'' at http://freshmeat.net/news/2000/08/05/965534399.html
Re: what is sufficient free memory?
Not to mention that unless you are having MILD traffic through the squid box, you probably want a box dedicated to just that. On Tue, 29 Aug 2000, Bulent Murtezaoglu wrote: | | Your biggest potential hog is squid. It maintains data structures in | memory and their size grows with your cache size. If anything causes | trashing that'll be it. The squid FAQ's give some back-of-envelope | calculations for this AFAIK. | | cheers, | | BM -- ___ _ __ _ __ /___ ___ /__ John Gonzalez/Net.Tech __ __ \ __ \ __/_ __ `__ \/ __ /_ ___/ MDC Computers/netMDC! _ / / / `__/ /_ / / / / / / /_/ / / /__ (505)439-0200/fax-437-3052 /_/ /_/\___/\__/ /_/ /_/ /_/\__,_/ \___/ http://www.netmdc.com [-[system info]---] 2:00pm up 110 days, 20:03, 4 users, load average: 0.04, 0.09, 0.23
dfme network driver
Hi: does anyone know if the dfme network driver is available during a debian 2.2 install? If so, dos it show up on the list of cards, and which hardare set should I get for the following items: I'll use a machine with udma66 drives, and the dfme network onboard. under RH Ihave to use insmod after install to make it work, but I'd like to do a net install of debian.
Re: what is sufficient free memory?
On Tue, 29 Aug 2000, Tamas TEVESZ wrote: > On Wed, 30 Aug 2000, Erik Peter P. Abella wrote: > > > 70 processes: 69 sleeping, 1 running, 0 zombie, 0 stopped > > CPU states: 0.1% user, 0.7% system, 0.0% nice, 99.0% idle > > Mem: 63124K av, 61296K used, 1828K free, 36880K shrd, 7712K buff > > Swap: 104380K av, 3128K used, 101252K free 35860K > > cached > > > > of that 61296 used k, 35860k is cache. filesystem cache. > > things happened so that linux uses the memory if it has. would it make > any sense having unused memory ? ok, it's cheap these days, but > not _that_ cheap... Why not that cheap? Less than half of his ram is being used for actual needed data. The rest is either free, in cache or in buffers which means the memory isn't even close to being stressed on the system. The 3MB in swap is just crap the system doesn't mind not having ;) I don't think the memory problem is as bad as only 1.8Mb free. -Nathan
Re: what is sufficient free memory?
Erik, linux will always use almost 100% of the memory (unless you have a BUTTLOAD of extra (ie. 512MB RAM)) for buffers and things like that. For example: skank:~# free total used free sharedbuffers cached Mem: 62956 61460 1496 15432 2064 12864 -/+ buffers/cache: 46532 16424 Swap: 102780 15476 87304 I only have 1396 bytes of memory free, and this machine has alot of processes running on it. What you really want to watch out for, is using alot of the swap. If your computer constantly dips heavily into the swap, it's time to add more memory. I recently increased the tasks that the above box needs to perform (mysql database for one, apache for two) so the memory requirements have no increased. Before, 64MB was more then sufficient. Now, it can handle it, but i would feel more comfortable with 128. I would suggest upgrading your box to 128MB at minimum as you are using a healthy portion of the swap, and being as how important that machine SEEMS (ie. the number of important tasks it performs) You may want to at least buy another stick of 128 as well, in case one of them goes bad, you'll have sufficient memory to run on in the mean time of a replacement. On Wed, 30 Aug 2000, Erik Peter P. Abella wrote: | Hello All, | | I had problems with the RAM (128MB SDRAM DIMM) of my server and the only | spare I had was (64MB SDRAM DIMM). | | Hear is what top says: | | 70 processes: 69 sleeping, 1 running, 0 zombie, 0 stopped | CPU states: 0.1% user, 0.7% system, 0.0% nice, 99.0% idle | Mem: 63124K av, 61296K used, 1828K free, 36880K shrd, 7712K buff | Swap: 104380K av, 3128K used, 101252K free 35860K | cached | | I'm a little worried that the harddisk might thrash as there's hardly | any more memory to spare (1828K free). This server is pretty much an | "all-in-one" box running DNS, squid, postfix, apache and radius (I have | 17 modems on it's cyclades). Any recommendations on the "bare-minimum" | RAM for this configuration? The suits at our purchasing division are | "stiff" at best so when I make the requesition for the replacement, I | need to know if I need to "demand" that they "immediately" cough-up the | 128mb (or more). | | | Thanks in advance, | | | Erik | | | -- | To UNSUBSCRIBE, email to [EMAIL PROTECTED] | with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] | | -- ___ _ __ _ __ /___ ___ /__ John Gonzalez/Net.Tech __ __ \ __ \ __/_ __ `__ \/ __ /_ ___/ MDC Computers/netMDC! _ / / / `__/ /_ / / / / / / /_/ / / /__ (505)439-0200/fax-437-3052 /_/ /_/\___/\__/ /_/ /_/ /_/\__,_/ \___/ http://www.netmdc.com [-[system info]---] 1:50pm up 110 days, 19:53, 4 users, load average: 0.10, 0.19, 0.37
Re: what is sufficient free memory?
On Wed, 30 Aug 2000, Erik Peter P. Abella wrote: > 70 processes: 69 sleeping, 1 running, 0 zombie, 0 stopped > CPU states: 0.1% user, 0.7% system, 0.0% nice, 99.0% idle > Mem: 63124K av, 61296K used, 1828K free, 36880K shrd, 7712K buff > Swap: 104380K av, 3128K used, 101252K free 35860K > cached > > I'm a little worried that the harddisk might thrash as there's hardly > any more memory to spare (1828K free). This server is pretty much an of that 61296 used k, 35860k is cache. filesystem cache. things happened so that linux uses the memory if it has. would it make any sense having unused memory ? ok, it's cheap these days, but not _that_ cheap... -- [-] ``And there are plenty of other innovative pieces of software such as Napster and ICQ.'' -- comment on ``Systems Software Research is Irrelevant'' at http://freshmeat.net/news/2000/08/05/965534399.html
what is sufficient free memory?
Your biggest potential hog is squid. It maintains data structures in memory and their size grows with your cache size. If anything causes trashing that'll be it. The squid FAQ's give some back-of-envelope calculations for this AFAIK. cheers, BM
what is sufficient free memory?
Hello All, I had problems with the RAM (128MB SDRAM DIMM) of my server and the only spare I had was (64MB SDRAM DIMM). Hear is what top says: 70 processes: 69 sleeping, 1 running, 0 zombie, 0 stopped CPU states: 0.1% user, 0.7% system, 0.0% nice, 99.0% idle Mem: 63124K av, 61296K used, 1828K free, 36880K shrd, 7712K buff Swap: 104380K av, 3128K used, 101252K free 35860K cached I'm a little worried that the harddisk might thrash as there's hardly any more memory to spare (1828K free). This server is pretty much an "all-in-one" box running DNS, squid, postfix, apache and radius (I have 17 modems on it's cyclades). Any recommendations on the "bare-minimum" RAM for this configuration? The suits at our purchasing division are "stiff" at best so when I make the requesition for the replacement, I need to know if I need to "demand" that they "immediately" cough-up the 128mb (or more). Thanks in advance, Erik
Re: what is sufficient free memory?
Hello again, Robert Davies wrote: > Why wait? > > Run the command vmstat, and observe how much is paged in /out, what is the > scan rate? That indicates how hard the page stealer is looking for pages it > can free off. vmstat 10 is usually a goodish, number but if you can run it > a long time, longer sample times is useful. OK, here goes: procs memoryswap io system cpu r b w swpd free buff cache si sobibo incs us sy id 0 0 0 3560 1644 31188 8608 0 0 4 1 10417 0 0 99 > If you still have the 128MB in the machine, you could force Linux to ignore > it, using a boot parameter. Nah, the machine won't boot unless I remove it. > With squid, you probably want to lower the mem cache down to about 1/4 of > physical RAM, if it's higher than the default. Depending on usage perhaps > less than 16MB would be a waste of RAM, your call! presumably via /etc/squid.conf ? > If the suits don't blink, then perhaps you could investigate interleaving > swap over a number of disk partions, by using mkswap, and swap entries in > fstab all set with a priority=5, instead of the defaults which don't > interleave. We'll see, I'm still waiting for a larger barracuda. It'd be nice if I can have both in one shabbang! for now I'm fairly confident that with the adjustment of squid, my swap partition can manage until I get the 128MB. Or won't that be enough? I'll soon find out and tell everyone how it goes. Mabuhay! Erik -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
ppp configuration PROBLEM
Help: I am new to Debian and am unabel to get internet applications running after connecting with wvdial?? Perhaps someone can advise me where I can enter my ISP's DNS address? Any help will be gratly appeciated. Jim Clawson __ Do You Yahoo!? Yahoo! Mail - Free email you can access from anywhere! http://mail.yahoo.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: what is sufficient free memory?
> Thanks for all the reponses. From most of the replies, can I gather that > I'll have to observe my how much is being swapped to determine whether I > should immediately "up" the RAM back to 128MB? (and pester the tight-wad > suits who'll approve the requesition) Why wait? Run the command vmstat, and observe how much is paged in /out, what is the scan rate? That indicates how hard the page stealer is looking for pages it can free off. vmstat 10 is usually a goodish, number but if you can run it a long time, longer sample times is useful. If you still have the 128MB in the machine, you could force Linux to ignore it, using a boot parameter. With lilo, that would mean an append line with 'mem=64M' in it, or enter it at the lilo boot prompt. boot: linux mem=64M That way you can actually trial your system out, at the cost of a reboot, to see the affect. With squid, you probably want to lower the mem cache down to about 1/4 of physical RAM, if it's higher than the default. Depending on usage perhaps less than 16MB would be a waste of RAM, your call! If the suits don't blink, then perhaps you could investigate interleaving swap over a number of disk partions, by using mkswap, and swap entries in fstab all set with a priority=5, instead of the defaults which don't interleave. Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Routing
On Tue, 29.08.00 09:48 -0700, Kevin wrote: > I've got my network on 10.0.0.0/24. The gateway is 10.0.0.1 and the > bridge/router is on 10.0.0.1. I need to setup a static route in the > gateway that says anything for 10.1.1.0/24 should use 10.0.0.1 as > its next hop. From my view I can't do it with normal route as it > will only take an interface as the destination. Any ideas? You want this? route add -net 10.1.1.0 netmask 255.255.255.0 gw 10.0.0.1 bye, -christian- -- Did You know that MicroSoft was named after Bill Gates' penis ?
Re: what is sufficient free memory?
Hello All, Thanks for all the reponses. From most of the replies, can I gather that I'll have to observe my how much is being swapped to determine whether I should immediately "up" the RAM back to 128MB? (and pester the tight-wad suits who'll approve the requesition) To paint a better picture, here's an entire top screen: 70 processes: 69 sleeping, 1 running, 0 zombie, 0 stopped CPU states: 0.1% user, 0.7% system, 0.0% nice, 99.0% idle Mem: 63124K av, 60764K used, 2360K free, 38700K shrd, 32216K buff Swap: 104380K av, 3572K used, 100808K free 7772K cached PID USER PRI NI SIZE RSS SHARE STAT LIB %CPU %MEM TIME COMMAND 1047 beyonder 10 0 1032 1032 820 R 0 0.7 1.6 0:00 top 717 root 1 0 1340 1288 1040 S 0 0.1 2.0 0:01 sshd 1 root 0 0 460 460 388 S 0 0.0 0.7 0:06 init 2 root 0 0 00 0 SW 0 0.0 0.0 0:00 kflushd 3 root 0 0 00 0 SW 0 0.0 0.0 0:00 kupdate 4 root 0 0 00 0 SW 0 0.0 0.0 0:00 kpiod 5 root 0 0 00 0 SW 0 0.0 0.0 0:00 kswapd 6 root -20 -20 00 0 SW< 0 0.0 0.0 0:00 mdrecoveryd 297 root 1 0 480 472 388 S 0 0.0 0.7 0:00 syslogd 308 root 0 0 684 672 316 S 0 0.0 1.0 0:00 klogd 324 root 1 0 548 548 480 S 0 0.0 0.8 0:00 crond 340 root 15 0 472 468 392 S 0 0.0 0.7 0:00 inetd 356 root 0 0 1556 1244 596 S 0 0.0 1.9 0:00 named 366 root 0 0 772 700 604 S 0 0.0 1.1 0:01 sshd 394 root 0 0 516 508 412 S 0 0.0 0.8 0:00 automount 452 root 2 0 688 688 548 S 0 0.0 1.0 0:00 master 458 postfix0 0 812 812 648 S 0 0.0 1.2 0:00 qmgr 470 root 0 0 400 384 324 S 0 0.0 0.6 0:00 gpm 486 root 0 0 1368 1364 1288 S 0 0.0 2.1 0:00 httpd 490 nobody 0 0 1220 1172 1044 S 0 0.0 1.8 0:00 httpd 491 nobody 0 0 1224 1224 1096 S 0 0.0 1.9 0:00 httpd 492 nobody 0 0 964 964 872 S 0 0.0 1.5 0:00 httpd 493 nobody 0 0 964 964 872 S 0 0.0 1.5 0:00 httpd 494 nobody 0 0 1220 1220 1096 S 0 0.0 1.9 0:00 httpd 495 nobody 0 0 1268 1268 1120 S 0 0.0 2.0 0:00 httpd The stats on the swapping have been constant for the past hour or so, but notable too is that it's 4am here in the Philippines and there aren't any subscribers who've dialed in and engaged squid as they surf. I guess I'll be able to get a better idea later on when my subscribers start pouring in. Mabuhay kayong lahat at debian! (long live you all and debian) Erik -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: what is sufficient free memory?
On Tue, 29 Aug 2000, Nathan wrote: > Why not that cheap? Less than half of his ram is being used for actual > needed data. The rest is either free, in cache or in buffers which means > the memory isn't even close to being stressed on the system. The 3MB in that's exactly what i'm talking about. the rest is being used as cache, just because it's there, and once it's there and it's not needed for ``real'' data storage, it's being used as cache. nothing's cheap enough just to lay around unused ;) -- [-] ``And there are plenty of other innovative pieces of software such as Napster and ICQ.'' -- comment on ``Systems Software Research is Irrelevant'' at http://freshmeat.net/news/2000/08/05/965534399.html -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: what is sufficient free memory?
Not to mention that unless you are having MILD traffic through the squid box, you probably want a box dedicated to just that. On Tue, 29 Aug 2000, Bulent Murtezaoglu wrote: | | Your biggest potential hog is squid. It maintains data structures in | memory and their size grows with your cache size. If anything causes | trashing that'll be it. The squid FAQ's give some back-of-envelope | calculations for this AFAIK. | | cheers, | | BM -- ___ _ __ _ __ /___ ___ /__ John Gonzalez/Net.Tech __ __ \ __ \ __/_ __ `__ \/ __ /_ ___/ MDC Computers/netMDC! _ / / / `__/ /_ / / / / / / /_/ / / /__ (505)439-0200/fax-437-3052 /_/ /_/\___/\__/ /_/ /_/ /_/\__,_/ \___/ http://www.netmdc.com [-[system info]---] 2:00pm up 110 days, 20:03, 4 users, load average: 0.04, 0.09, 0.23 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
dfme network driver
Hi: does anyone know if the dfme network driver is available during a debian 2.2 install? If so, dos it show up on the list of cards, and which hardare set should I get for the following items: I'll use a machine with udma66 drives, and the dfme network onboard. under RH Ihave to use insmod after install to make it work, but I'd like to do a net install of debian. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: what is sufficient free memory?
On Tue, 29 Aug 2000, Tamas TEVESZ wrote: > On Wed, 30 Aug 2000, Erik Peter P. Abella wrote: > > > 70 processes: 69 sleeping, 1 running, 0 zombie, 0 stopped > > CPU states: 0.1% user, 0.7% system, 0.0% nice, 99.0% idle > > Mem: 63124K av, 61296K used, 1828K free, 36880K shrd, 7712K buff > > Swap: 104380K av, 3128K used, 101252K free 35860K > > cached > > > > of that 61296 used k, 35860k is cache. filesystem cache. > > things happened so that linux uses the memory if it has. would it make > any sense having unused memory ? ok, it's cheap these days, but > not _that_ cheap... Why not that cheap? Less than half of his ram is being used for actual needed data. The rest is either free, in cache or in buffers which means the memory isn't even close to being stressed on the system. The 3MB in swap is just crap the system doesn't mind not having ;) I don't think the memory problem is as bad as only 1.8Mb free. -Nathan -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: what is sufficient free memory?
Erik, linux will always use almost 100% of the memory (unless you have a BUTTLOAD of extra (ie. 512MB RAM)) for buffers and things like that. For example: skank:~# free total used free sharedbuffers cached Mem: 62956 61460 1496 15432 2064 12864 -/+ buffers/cache: 46532 16424 Swap: 102780 15476 87304 I only have 1396 bytes of memory free, and this machine has alot of processes running on it. What you really want to watch out for, is using alot of the swap. If your computer constantly dips heavily into the swap, it's time to add more memory. I recently increased the tasks that the above box needs to perform (mysql database for one, apache for two) so the memory requirements have no increased. Before, 64MB was more then sufficient. Now, it can handle it, but i would feel more comfortable with 128. I would suggest upgrading your box to 128MB at minimum as you are using a healthy portion of the swap, and being as how important that machine SEEMS (ie. the number of important tasks it performs) You may want to at least buy another stick of 128 as well, in case one of them goes bad, you'll have sufficient memory to run on in the mean time of a replacement. On Wed, 30 Aug 2000, Erik Peter P. Abella wrote: | Hello All, | | I had problems with the RAM (128MB SDRAM DIMM) of my server and the only | spare I had was (64MB SDRAM DIMM). | | Hear is what top says: | | 70 processes: 69 sleeping, 1 running, 0 zombie, 0 stopped | CPU states: 0.1% user, 0.7% system, 0.0% nice, 99.0% idle | Mem: 63124K av, 61296K used, 1828K free, 36880K shrd, 7712K buff | Swap: 104380K av, 3128K used, 101252K free 35860K | cached | | I'm a little worried that the harddisk might thrash as there's hardly | any more memory to spare (1828K free). This server is pretty much an | "all-in-one" box running DNS, squid, postfix, apache and radius (I have | 17 modems on it's cyclades). Any recommendations on the "bare-minimum" | RAM for this configuration? The suits at our purchasing division are | "stiff" at best so when I make the requesition for the replacement, I | need to know if I need to "demand" that they "immediately" cough-up the | 128mb (or more). | | | Thanks in advance, | | | Erik | | | -- | To UNSUBSCRIBE, email to [EMAIL PROTECTED] | with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] | | -- ___ _ __ _ __ /___ ___ /__ John Gonzalez/Net.Tech __ __ \ __ \ __/_ __ `__ \/ __ /_ ___/ MDC Computers/netMDC! _ / / / `__/ /_ / / / / / / /_/ / / /__ (505)439-0200/fax-437-3052 /_/ /_/\___/\__/ /_/ /_/ /_/\__,_/ \___/ http://www.netmdc.com [-[system info]---] 1:50pm up 110 days, 19:53, 4 users, load average: 0.10, 0.19, 0.37 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: what is sufficient free memory?
On Wed, 30 Aug 2000, Erik Peter P. Abella wrote: > 70 processes: 69 sleeping, 1 running, 0 zombie, 0 stopped > CPU states: 0.1% user, 0.7% system, 0.0% nice, 99.0% idle > Mem: 63124K av, 61296K used, 1828K free, 36880K shrd, 7712K buff > Swap: 104380K av, 3128K used, 101252K free 35860K > cached > > I'm a little worried that the harddisk might thrash as there's hardly > any more memory to spare (1828K free). This server is pretty much an of that 61296 used k, 35860k is cache. filesystem cache. things happened so that linux uses the memory if it has. would it make any sense having unused memory ? ok, it's cheap these days, but not _that_ cheap... -- [-] ``And there are plenty of other innovative pieces of software such as Napster and ICQ.'' -- comment on ``Systems Software Research is Irrelevant'' at http://freshmeat.net/news/2000/08/05/965534399.html -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
what is sufficient free memory?
Your biggest potential hog is squid. It maintains data structures in memory and their size grows with your cache size. If anything causes trashing that'll be it. The squid FAQ's give some back-of-envelope calculations for this AFAIK. cheers, BM -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
what is sufficient free memory?
Hello All, I had problems with the RAM (128MB SDRAM DIMM) of my server and the only spare I had was (64MB SDRAM DIMM). Hear is what top says: 70 processes: 69 sleeping, 1 running, 0 zombie, 0 stopped CPU states: 0.1% user, 0.7% system, 0.0% nice, 99.0% idle Mem: 63124K av, 61296K used, 1828K free, 36880K shrd, 7712K buff Swap: 104380K av, 3128K used, 101252K free 35860K cached I'm a little worried that the harddisk might thrash as there's hardly any more memory to spare (1828K free). This server is pretty much an "all-in-one" box running DNS, squid, postfix, apache and radius (I have 17 modems on it's cyclades). Any recommendations on the "bare-minimum" RAM for this configuration? The suits at our purchasing division are "stiff" at best so when I make the requesition for the replacement, I need to know if I need to "demand" that they "immediately" cough-up the 128mb (or more). Thanks in advance, Erik -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Routing
I've got my network on 10.0.0.0/24. The gateway is 10.0.0.1 and the bridge/router is on 10.0.0.1. I need to setup a static route in the gateway that says anything for 10.1.1.0/24 should use 10.0.0.1 as its next hop. From my view I can't do it with normal route as it will only take an interface as the destination. Any ideas? -- Kevin - [EMAIL PROTECTED]
Re: Routing
On Tue, 29.08.00 09:48 -0700, Kevin wrote: > I've got my network on 10.0.0.0/24. The gateway is 10.0.0.1 and the > bridge/router is on 10.0.0.1. I need to setup a static route in the > gateway that says anything for 10.1.1.0/24 should use 10.0.0.1 as > its next hop. From my view I can't do it with normal route as it > will only take an interface as the destination. Any ideas? You want this? route add -net 10.1.1.0 netmask 255.255.255.0 gw 10.0.0.1 bye, -christian- -- Did You know that MicroSoft was named after Bill Gates' penis ? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Web crawler engine.
There are lot of web search engines around, i mean like altavista's google's infoseek's etc.. What does one do when she wants to build one herself. I mean what do you people use when you build local search engine ( city, locality, my needs are about 100 000 pages?webs? for the beggining). I've been looking at some open-source engines like htdig, swish-e, swish++, udmsearch and the like.. all of them have some flaws which make them unusable - htdig keeps it's data in db-files not in sql, same goes for swishes, and udmsearch is written in highly undocumented and hairy c. byuing services from altavista seems to be out of my league, google would be nice, but building local service I need to make it index only local things, and that can be done only be using long list of ip-networks. So I need to rewrite indexing part of engine to implement that. Anyone doing sth like that? any comments and help appreciated. regards, Eyck
reiserfs & databases.
AFAIK reiserfs is about keeping files (blocks) in b-trees, and DBMS keep their data in a bunch of files, which are accessed directly (non-sequential access). So I figured that reiserfs would be great for keeping DBMS's data on it. but, there are some commercial databases which keep their data directly on partitions ( this should be much better then any *fs including reiserfs) and the weird part is that that direct-partition instalation scheme seems to be a little bit slower that fs-based in benchmarks. And this means that I'm missing something here, what is it that I haven't thought about, anyone, any comments on this? regards, Eyck
Routing
I've got my network on 10.0.0.0/24. The gateway is 10.0.0.1 and the bridge/router is on 10.0.0.1. I need to setup a static route in the gateway that says anything for 10.1.1.0/24 should use 10.0.0.1 as its next hop. From my view I can't do it with normal route as it will only take an interface as the destination. Any ideas? -- Kevin - [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
apache & ssl & certificates,
Hello, I am having a little problem with ?x.509? certificates and virtual domains, it seems like name (not ip) of host is being put in certificate, so when I sign my cet as host.net, then when someone connects to virtual.host.net she gets 'hosts name differ from certificate' or sth error. Same goes to ssl tunels for pop3. What can I do?
Web crawler engine.
There are lot of web search engines around, i mean like altavista's google's infoseek's etc.. What does one do when she wants to build one herself. I mean what do you people use when you build local search engine ( city, locality, my needs are about 100 000 pages?webs? for the beggining). I've been looking at some open-source engines like htdig, swish-e, swish++, udmsearch and the like.. all of them have some flaws which make them unusable - htdig keeps it's data in db-files not in sql, same goes for swishes, and udmsearch is written in highly undocumented and hairy c. byuing services from altavista seems to be out of my league, google would be nice, but building local service I need to make it index only local things, and that can be done only be using long list of ip-networks. So I need to rewrite indexing part of engine to implement that. Anyone doing sth like that? any comments and help appreciated. regards, Eyck -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
reiserfs & databases.
AFAIK reiserfs is about keeping files (blocks) in b-trees, and DBMS keep their data in a bunch of files, which are accessed directly (non-sequential access). So I figured that reiserfs would be great for keeping DBMS's data on it. but, there are some commercial databases which keep their data directly on partitions ( this should be much better then any *fs including reiserfs) and the weird part is that that direct-partition instalation scheme seems to be a little bit slower that fs-based in benchmarks. And this means that I'm missing something here, what is it that I haven't thought about, anyone, any comments on this? regards, Eyck -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
apache & ssl & certificates,
Hello, I am having a little problem with ?x.509? certificates and virtual domains, it seems like name (not ip) of host is being put in certificate, so when I sign my cet as host.net, then when someone connects to virtual.host.net she gets 'hosts name differ from certificate' or sth error. Same goes to ssl tunels for pop3. What can I do? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
