Re: [SLUG] Swap file performance tuning ?
Rod Butcher wrote: Sluggers, I have 1 gig ram and 4 gig Swap on a Sata drive... performance when using memory-hogs like Sweep audio editor is great while it's using ram, but slows to a crawl once it starts using swap... not just Sweep but any app. Is this normal, or can I tune this in some way ? (I've spent my hardware budget for the next 3 years so no more ram). Reduce your SWAP to between 1 X 1 to 1 X 2 RAM. Ideally 1 X 1 RAM-to-swap. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
RE: [SLUG] Swap file performance tuning ?
Reduce your SWAP to between 1 X 1 to 1 X 2 RAM. Ideally 1 X 1 RAM-to-swap. I know that's the standard recommendation one sees for creating swap partitions, but I'm intrigued as to the reasons for your suggestion. How does reducing the swap:ram ratio improve performance? I expect the kernel would be conservative about using swap irrespective of how much there is. 1:1 RAM-to-swap is not at all ideal if your applications need 2Gb of memory and you've only got 512Mb swap to go with your 512Mb RAM and the machine crashes. My understanding is that having plenty of swap space isn't a bad thing if you can spare the space and are likely to make use of it. I'd like to know whether that's a misconception that actually degrades performance, and why. Cheers, - Rog -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Re: Can't reboot: cannot open /mnt/etc/fstab: no such file
Apologies for the alarm. I've managed to fix it - at least until the next reboot (sickly grin). Cheers, Adam Bogacki, [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Email filtering in Evolution
Is anyone having problems with email filtering not working in Evolution? No matter how I set up the filters I just cannot get Evolution to flick the emails into the correct folders back on the IMAP server, or does Evolution only flick them into its virtual folders? Yes, I have RTFM. -- Howard. LANNet Computing Associates; Your Linux people http://www.lannetlinux.com -- When you just want a system that works, you choose Linux; when you want a system that just works, you choose Microsoft. -- Flatter government, not fatter government; Get rid of the Australian states. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Swap file performance tuning ?
Roger Barnes wrote: Reduce your SWAP to between 1 X 1 to 1 X 2 RAM. Ideally 1 X 1 RAM-to-swap. I know that's the standard recommendation one sees for creating swap partitions, but I'm intrigued as to the reasons for your suggestion. How does reducing the swap:ram ratio improve performance? I expect the kernel would be conservative about using swap irrespective of how much there is. 1:1 RAM-to-swap is not at all ideal if your applications need 2Gb of memory and you've only got 512Mb swap to go with your 512Mb RAM and the machine crashes. My understanding is that having plenty of swap space isn't a bad thing if you can spare the space and are likely to make use of it. I'd like to know whether that's a misconception that actually degrades performance, and why. I'm not the original person who suggested this but I lean towards supporting this suggestion. I too used to think that too large swap area doesn't hurt, but the explenation I heard, which sounds reasonable to me though I never got around to test it, is that lots of swap also means lots of book-keeping - the kernel takes more time to manage the free swap pages and scan through them. Of course if you reduce your swap to a level that your application can't handle your data then it's a problem. Cheers, - Rog Cheers, --Amos -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Swap file performance tuning ?
Roger Barnes wrote: I know that's the standard recommendation one sees for creating swap partitions, but I'm intrigued as to the reasons for your suggestion. How does reducing the swap:ram ratio improve performance? I expect the kernel would be conservative about using swap irrespective of how much there is. When SWAP space is utilised by the system it is done so by writing snapshots of memory called pages. These pages are written NOT IN contigous locations within the swap area and as a result when these pages are re-called the system has to work harder to locate these pages on top of the overhead of the SLOW access of physical disk. This is exacerbated by increasing the physical area over which the disk heads traversed. This is further exacerbated if you SWAP SPACE resides in the same disk as your busy files. The system is always looking for data in an area where data is fragmented. As the area of the SWAP space becomes bigger this problem is even more exacerbated. This performance loss with big SWAP space becomes apparent as DRAM becomes faster-and-faster but disk storage technology is not keeping up proportion-wise. Some 20 years ago it does not matter even if one has large swap space because RAM speeds are not so much faster than disk access speed. I just say 1-to-1 mem-to-swap ratio because that is the rule-of-thumb recommendations. When I tune systems under my control I try to allocate as little SWAP as I can get away with without making a system thrash. If you do not have software with memory leak you will almost certainly have a buzzing system. Secondly, by default i386 systems, which I assume your systems is, will only be able to utilize a SWAP SPACE of NOT OVER 2GB due to file size limitations. Of course you can modify kernel params to say LARGE-FILE is supported so you may utilise the full 4GB when required. But why would you need 4GB of swap when you have 1GB of memory. My understanding is that having plenty of swap space isn't a bad thing if you can spare the space and are likely to make use of it. I'd like to know whether that's a misconception that actually degrades performance, and why. Yes, this is a misconception. When you want speed, do not use SWAP space. When you are after many loads or users use SWAP space. If you want in-between then tune SWAP space. I have actually experienced what you did and that is having large SWAP space but discovering the system became slower for no benefit. It is a trade-off . -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Swap file performance tuning ?
On Mon, 2004-10-25 at 13:04 +0200, [EMAIL PROTECTED] wrote: I too used to think that too large swap area doesn't hurt, but the explenation I heard, which sounds reasonable to me though I never got around to test it, is that lots of swap also means lots of book-keeping - the kernel takes more time to manage the free swap pages and scan through them. Of course if you reduce your swap to a level that your application can't handle your data then it's a problem. IIRC there used to be problems with bad-O algorithms in the page table, but in 2.6 they are fixed as part of the scalability work. Certainly there is extra overhead in non-swappable kernel memory in maintaining the larger page table. Thats fairly minimal (~32 bytes per 4Kb on ix86, IIRC). Rob -- GPG key available at: http://www.robertcollins.net/keys.txt. signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Swap file performance tuning ?
...Houston... uh... we have a problem.. swap file is uh.. full... sheeit, I told you to make it bigger... whoa.. watch that crater... --- Brought to you by a thunderbird, penguin, gnu and a camel Roger Barnes wrote: Reduce your SWAP to between 1 X 1 to 1 X 2 RAM. Ideally 1 X 1 RAM-to-swap. I know that's the standard recommendation one sees for creating swap partitions, but I'm intrigued as to the reasons for your suggestion. How does reducing the swap:ram ratio improve performance? I expect the kernel would be conservative about using swap irrespective of how much there is. 1:1 RAM-to-swap is not at all ideal if your applications need 2Gb of memory and you've only got 512Mb swap to go with your 512Mb RAM and the machine crashes. My understanding is that having plenty of swap space isn't a bad thing if you can spare the space and are likely to make use of it. I'd like to know whether that's a misconception that actually degrades performance, and why. Cheers, - Rog -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Swap file performance tuning ?
Robert Collins wrote: On Mon, 2004-10-25 at 13:04 +0200, [EMAIL PROTECTED] wrote: I too used to think that too large swap area doesn't hurt, but the explenation I heard, which sounds reasonable to me though I never got around to test it, is that lots of swap also means lots of book-keeping - the kernel takes more time to manage the free swap pages and scan through them. Of course if you reduce your swap to a level that your application can't handle your data then it's a problem. IIRC there used to be problems with bad-O algorithms in the page table, but in 2.6 they are fixed as part of the scalability work. Certainly there is extra overhead in non-swappable kernel memory in maintaining the larger page table. Thats fairly minimal (~32 bytes per 4Kb on ix86, IIRC). The overhead I though about was more about the CPU time required to sift through the records rather than the amount of memory it takes. I haven't heard about the 2.6 improvments you talk about, so what you are saying is that large swap is not such a bad idea under 2.6? Cheers, --Amos -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Update on SLUG Goonellabah roadtrip
On Mon, 25 Oct 2004 11:35:55 +1000 Jan Schmidt [EMAIL PROTECTED] wrote: I'm postponing the trip until a later date. One of the guys that was coming along won't be able to make it, and the SLUG meeting is on Friday as well, so I think we'll get a bigger group and have more fun if we make it a later date. Sorry to hear this, but the more the merrier. :-) Hopefully Lindsay will be well enough to join us when we go :) Chemo is coming along well. Feel pretty off most of the time but i'm making a speedy recovery. Roadtripping will be a welcome relief. Cheers, Lindsay -- http://www.asymmetrics.net/~auxesis/ -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] commonwealth bank netbanking stopped support linux?
Netbank works fine, all you need is javascript support basically. So far it seems to work in any browser I have tried. I usually use it in konqueror (khtml) and it's fine. The address I always use to connect to is http://www.netbank.commbank.com.au/ NOTE: No https here it will redirect to a https:// url by itself. This box has a history of difficulties with Netbank. Netbank works fine for me (now) with Mozilla 1.6 and I have java enabled as a plug-in and disabled in 'Advanced' The ComBank site works better that way with this box, I don't know, just noticed that along the way. ie in Mozilla menu - edit | preferences | Advanced | Scripts and Plugins then in Scripts and Plug-ins dialog, tick box against Navigator I go in through the Combank front door - http://www.commbank.com.au/default.asp and click on Netbank login button. Ocassionally, in the past, the front door approach didn't work as it returned the browser back to the 'front door' without opening up the Netbank page, so go through - http://www.commbank.com.au/Netbank/intro/ go right to the end of the page and click on highlighted 'start' the Mozilla and java packages: [EMAIL PROTECTED]:~$ apt-show-versions | grep mozilla mozilla-browser/testing upgradeable from 2:1.6-5 to 2:1.7.3-5 mozilla/testing upgradeable from 2:1.6-5 to 2:1.7.3-5 mozilla-mailnews/testing upgradeable from 2:1.6-5 to 2:1.7.3-5 mozilla-psm/testing upgradeable from 2:1.6-5 to 2:1.7.3-5 mozilla-xft 2:1.6-5 installed: No available version in archive [EMAIL PROTECTED]:~$ apt-show-versions | grep java java-virtual-machine-dummy/stable uptodate 0.3 java2-common/testing uptodate 1 java-common/testing uptodate 0.22 regards Russell -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Swap file performance tuning ?
The stuff at http://www.fiveanddime.net/ss/swap.htm about configuring multiple parallel swap files looks promising... I have 2 SATA drives that I could put 2 gig on each... but I don't know its author, and the page is undated... so before I destroy my system I'd like to run this past yous.. (and yes, I do need all that memory). My current fstab swap setting (by the Mandrake install) :- /dev/hda5 swap swap defaults 0 0 question - what are the defaults ? Recommended setting in the article mentioned above :- /dev/hdc5 noneswapsw,pri=30 0 /dev/hda5 noneswapsw,pri=30 0 This should cause the system to use both in parallel because they have the same priority, and apparently go like blazes. The sw,pri=3 seems OK according to man swapon, but it's dated 1995. Do these options look sane ? thanks Rod --- Brought to you by a thunderbird, penguin, gnu and a camel Roger Barnes wrote: Hiya Rod, 1. What's the difference between buffer-cache reads and buffered disk reads ? I have 3 disks and the figures vary.. with the cache read mb figues approx. 10 times those of the disk reads. Which figure is more important re. swap ? My understanding (which could be incorrect) is that the cached reads are testing the speed of the small bit of volatile cache on the disk (it's fast, but it's not the magnetic bit where the data lives permanently). In reality, throughput is going to be closer to the buffered disk read speed. For swap, or any purpose really, you really should benchmark something realistic and if you're that keen, you might be able to find some applications that do that. The hdparm figures are raw read speeds with and without using the disk cache respectively. For improving things in your situation, I'd suggest you just look at the buffered disk read speed, then see if that number improves with any tweaking you do. 2. Are you implying I can have multiple swap files actyive simultaneously ? Certainly. Apparently linux is reasonably clever about making good use of it too. A link that I just found with a bit of googling seems worthy of a good read ... http://www.fiveanddime.net/ss/swap.htm Cheers, - Rog . -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] filters in Evolution not triggering
Howard, Yes, you are not alone. Filters work in Evolution 1.4, but do not seem to be automatically invoked. I find I have to ctrl-A to Select All, then ctrl-Y to apply filters. Once manually initiated, the filter rules work fine, but they do not seem to trigger automatically. The setup has Evolution 1.4 on RHEL via MAPI from Exchange. (really) HTH Brian -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Swap file performance tuning ?
On Mon, 2004-10-25 at 13:34 +0200, [EMAIL PROTECTED] wrote: The overhead I though about was more about the CPU time required to sift through the records rather than the amount of memory it takes. Last I looked in detail, its a three-level tree lookup - pretty much constant regardless of swap size. I haven't heard about the 2.6 improvments you talk about, so what you are saying is that large swap is not such a bad idea under 2.6? Yup. -- GPG key available at: http://www.robertcollins.net/keys.txt. signature.asc Description: This is a digitally signed message part -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Re: Can't reboot: cannot open /mnt/etc/fstab: no such file
Apparently, _Adam Bogacki_, on 25/10/04 06:31,typed: Apologies for the alarm. I've managed to fix it - at least until the next reboot (sickly grin). Cheers, Adam Bogacki, [EMAIL PROTECTED] I think you should also mention the steps you took to fix it so that others who hit this post in their search may find it useful. -HS -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Re: [nylug-talk] question to mutt users
Nothing is wrong with using fetchmail. This is just what I use. -- Dave Peters On Mon, Oct 25, 2004 at 07:53:43AM +0200, [EMAIL PROTECTED] wrote: Dave wrote: Hi, I think there are really two things that you may be trying to accomplish. Getting mail - which mutt doesn't do and. . . Sorting mail - which mutt doesn't do. I use a quick perl script to get my mail and pass it to procmail to do the sorting. I just set this up as a cron job. Here is the perl script that I use . . . =-=- #!/usr/bin/perl use Net::POP3; #Constructors $username=dave; $password=password; $pop = Net::POP3-new('mail.myhost.net', Timeout = 60); $msg_count = $pop-login($username, $password); if ($msg_count 0) { my $msgnums = $pop-list; # hashref of msgnum = size my $ping = $pop-ping( $username ); # print You have mail - $msg_count\n; foreach my $msgnum (keys %$msgnums) { #print getting message $msgnum . . . \n; my $msg = $pop-get($msgnum); open ( PROCMAIL, |/usr/bin/procmail ); print PROCMAIL @$msg; close ( PROCMAIL ); $pop-delete($msgnum); } } $pop-quit; =-=- My procmailrc dumps email through spamassasin then off to my /mail/ directory. If anyone has suggestions on the above script I wouldn't mind making it better. What's wrong with using fetchmail? Hope this helps Cheers, --Amos -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Re: Can't reboot: cannot open /mnt/etc/fstab: no such file
H. S. wrote: Apparently, _Adam Bogacki_, on 25/10/04 06:31,typed: Apologies for the alarm. I've managed to fix it - at least until the next reboot (sickly grin). Cheers, Adam Bogacki, [EMAIL PROTECTED] I think you should also mention the steps you took to fix it so that others who hit this post in their search may find it useful. -HS Thanks. I noticed the error message mentioned 'timeout' so I thought I'd let the system work it through. I watched the first half of 'Pearl Harbour' (not recommended) and came back to find the logon screen waiting for me. It's twice rebooted without problems since then. I put it down to software teething problems and the unexpected arrival of an upgrade which remembers the settings of the previous user session. It threw me for a while, but I'm impressed. Adam Bogacki, [EMAIL PROTECTED] -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Re: Can't reboot: cannot open /mnt/etc/fstab: no such file
quote who=Adam Bogacki I noticed the error message mentioned 'timeout' so I thought I'd let the system work it through. I watched the first half of 'Pearl Harbour' (not recommended) and came back to find the logon screen waiting for me. It's twice rebooted without problems since then. I put it down to software teething problems and the unexpected arrival of an upgrade which remembers the settings of the previous user session. It threw me for a while, but I'm impressed. Urk, personally, I'd be more worried than that. I've normally only seen timeouts like that when a) The hard disk was about to die b) The disk geometry was configured incorrectly, and the kernel was trying to seek past the end of the disk c) A bug in the kernel was simulating one of the first two. J. -- Jan Schmidt [EMAIL PROTECTED] ENOSIG -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Swap file performance tuning ?
Rod Butcher wrote: Sluggers, I have 1 gig ram and 4 gig Swap on a Sata drive... performance when using memory-hogs like Sweep audio editor is great while it's using ram, but slows to a crawl once it starts using swap... not just Sweep but any app. Is this normal, or can I tune this in some way ? (I've spent my hardware budget for the next 3 years so no more ram). 1) Don't swap It is slow, that's the nature of the beast. 2) Put it on a partition, not in a file. You don't want the filesystem doing wierd performance-reducing stuff. Make sure you are using the most recent swap format. 3) Put the partition on fast disk A disk of its own is good, since then the disk head movement from paging doesn't drop your application performance, and vice-versa. Alternatively, a partition in the center of the disk, so the head never needs to move more than 0.5 of a disk. 4) Know the limits I think the max swap partition size is 2GB, although you can have multiple. I can't see why you'd set up a smaller partition in this age of huge disks. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] dialup Win client/Woody
I'm trying to dial into my Woody server from a WinMe box, something I've never had the displeasure to do before. It works perfectly using PPP from all my Macintoshes. I get this in the authlog: mgetty[11769]: init chat failed, exiting...: Invalid argument mgetty[11769]: failed in mg_init_data Is there a difference between Windows and Macintosh ppp clients? Windows is one of life's mysteries as far as I am concerned :( thanks.. . David. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] dialup Win client/Woody
check the settings in your windows me box, its probably trying to use ms-chap or something bizarre good rule of thumb is to turn everything off ;) depending on your macos version, likely its just the same ppp software as on your linux machine. Dean On Tue, October 26, 2004 10:56 am, David said: I'm trying to dial into my Woody server from a WinMe box, something I've never had the displeasure to do before. It works perfectly using PPP from all my Macintoshes. I get this in the authlog: mgetty[11769]: init chat failed, exiting...: Invalid argument mgetty[11769]: failed in mg_init_data Is there a difference between Windows and Macintosh ppp clients? Windows is one of life's mysteries as far as I am concerned :( thanks.. . David. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Weblog software recommendations
I recommend checking out hobix (http://hobix.com/) by whytheluckystiff (http://whytheluckystiff.net/), even if you don't intend to use it. It's a blog written entirely in ruby and based on YAML, Textile, and WEBricks. Don't be afraid of his crazy antics/ramblings...he's actually a very cool ruby coder and you might actually learn a thing or two after laughing yourself silly. Who said setting up a blog can't be hilarious? ;) Also, a little off topic, if you're after a cool wiki written in ruby I recommend checking out instiki (http://www.instiki.org/show/HomePage). cya, nullobject On Thu, 21 Oct 2004 21:15:31 +0200, Mary Gardiner [EMAIL PROTECTED] wrote: On Thu, Oct 21, 2004, Mary Gardiner wrote: Have you ever used Pivot? I'm asking because I want to know whether it is capable of working with vhosting. It seems to really expect that all the weblogs are going to live 'near' its own directory, not in totally different domains. OK, a quick review of Pivot: 1. It's quite pretty and seems reasonable intuitive at first. 2. Making any changes at all to the templates require that an administrator moves template pages around for you. 3. Its multiple weblog handling is rather bizarre. #2 is a severe handicap: several of my users know basic HTML and like to photoshop up some images and whack them in, and all the weblogs look different, I don't want to have to maintain an enormous directory with template files for 10 (yes, literally) different weblogs. To elaborate on #3: at no point do you ever make an entry and say I want it to go in Mary's weblog or I want it to go in June's weblog. Instead, you write and entry and say this entry is in category Music. Elsewhere, you decide that all entries in category Music get posted in both Mary's weblog and Beth's weblog, but not in June's weblog. So, when you make a new entry, you choose its category, and then the pre-existing configuration chooses the weblog(s) it goes in. While I can imagine a use-case for this configuration (it would be OK for a single user who had several weblogs targetted at slightly different readers, so they could post My Favourite Drug categorised entries to the Friends weblog and the Dealers weblog but not the Family one) it's pretty much a deal-breaker for me: training users that to get an entry to appear on your personal weblog, you need to mark it as being in the mary_entry category is very much outside the way they currently use the weblogs. Categories, like subdirectories and mail folders, are a pretty sophisticated way of organising data. -Mary -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
[SLUG] Banks colluding with Microsoft ?
Sluggers, I started off having to spoof my Firefox browser as Moz 1 Win98 to do St George internet banking. That stopped working so I used IE 5 Mac. That now crashes and I have to use IE6 Win XP. There is as pattern here .. and in the US i believe it would warrant antitrust / racketeering / cartel / wirefraud investigation. Anybody got any real facts on this ? regards Rod --- Brought to you by a thunderbird, penguin, gnu and a camel -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Swap file performance tuning ?
Rod Butcher [EMAIL PROTECTED] writes: Sluggers, I have 1 gig ram and 4 gig Swap on a Sata drive... performance when using memory-hogs like Sweep audio editor is great while it's using ram, but slows to a crawl once it starts using swap... not just Sweep but any app. Is this normal, or can I tune this in some way ? (I've spent my hardware budget for the next 3 years so no more ram). thanks My laptop needs to use swap, and there is a noticeable difference in performance between different kernel versions as developers have directly or indirectly changed the behaviour of swap. There's a knob which you can play with in /proc/sys/vm/swappiness, there's probably other things you can tweak, too. If you like patching and compiling your own kernels, try the con kolivas tree, http://members.optusnet.com.au/ckolivas/kernel/. It's designed to improve responsiveness on desktops. I have no idea how it will go with your work load, though, it sounds like you need to swap, and no tuning or mucking around is going to have much effect. In my case, there isn't one application that needs loads of memory, just lots of applications that need a little, so the balance of the size of filesystem cache/buffers, and what, when and how much to swap something out has an effect. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Banks colluding with Microsoft ?
On Tue, 2004-10-26 at 14:50, Rod Butcher wrote: Sluggers, I started off having to spoof my Firefox browser as Moz 1 Win98 to do St George internet banking. That stopped working so I used IE 5 Mac. That now crashes and I have to use IE6 Win XP. There is as pattern here .. and in the US i believe it would warrant antitrust / racketeering / cartel / wirefraud investigation. Anybody got any real facts on this ? Ockams Razor (paraphrased): Never ascribe to conspiracy what can be easily attributed to stupidity. regards Rod --- Brought to you by a thunderbird, penguin, gnu and a camel -- Howard. LANNet Computing Associates; Your Linux people http://www.lannetlinux.com -- When you just want a system that works, you choose Linux; when you want a system that just works, you choose Microsoft. -- Flatter government, not fatter government; Get rid of the Australian states. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Banks colluding with Microsoft ?
I used do do this stuff for various clients including the oil industry.. and managed to keep interactive websites (orders, registrations, remote website updates..) operable and bulletproof with complex javascript etc. for the current AND previous versions of IE, NS, Opera, Knoqueror, on Win, Linux, Mac... all by myself. If they didn't work they could cause a shareholderor director to ring up and give me or the oil company an earful, which was a suboptimal outcome. So I struggle to belive that these genius graduate software engineers struggle with it. Trouble is I've forgotton all the technical stuff now and do other things. I'll let it drop here. cheers Rod --- Brought to you by a thunderbird, penguin, gnu and a camel Roger Barnes wrote: Verging on OT, so posting to chat instead ... Sluggers, I started off having to spoof my Firefox browser as Moz 1 Win98 to do St George internet banking. That stopped working so I used IE 5 Mac. That now crashes and I have to use IE6 Win XP. There is as pattern here .. and in the US i believe it would warrant antitrust / racketeering / cartel / wirefraud investigation. You better put an extra layer on that tin-foil hat. :p I'm all for a good conspiracy theory, but I think the tendency of financial institutions to block based on User-agent is a frail attempt to cover their butts by only officially supporting browsers that they have tested their site against. Net banking applications go through a heck of a lot more testing than most web sites, and reliably supporting 2-3 times as many browsers for a small fraction of users is firmly wedged in their too hard and not worth it basket. Neither the banks, nor M$ are involved with Firefox/Mozilla. Irrespective of the remote host's behaviour, your browser shouldn't be crashing. Have you filed a bug report? Keep applying the pressure by complaining about the poor service, or take it the appropriate industry ombudsman and the ACCC. - Rog . -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Banks colluding with Microsoft ?
; I used do do this stuff for various clients including the oil industry.. ; and managed to keep interactive websites (orders, registrations, remote ; website updates..) operable and bulletproof with complex javascript etc. ; for the current AND previous versions of IE, NS, Opera, Knoqueror, on ; Win, Linux, Mac... all by myself. If they didn't work they could cause a ; shareholderor director to ring up and give me or the oil company an ; earful, which was a suboptimal outcome. So I struggle to belive that ; these genius graduate software engineers struggle with it. Trouble is ; I've forgotton all the technical stuff now and do other things. I'll let ; it drop here. ok -- this is ridiculous. I've just tried st george with firefox, Error: document.bhjvmd.getJavaVendor is not a function Source File: https://ibank.stgeorge.com.au/html/redirect.asp Line: 11 source: HTMLHEAD !-- These scripts are provided under the terms of the BrowserHawk license agreement and may not be copied or used otherwise. [7, 0, 1, 0 Enterprise] See cyscape.com for details. Copyright (C) 1999-2003 cyScape, Inc. All rights reserved. -- noscriptmeta http-equiv=refresh content=0; url=/html/redirect.asp?bhjs=0/noscript script language=JavaScript !-- function bhawkTest() { var bhjv = escape('Java N/A'); if (document.bhjvmd) bhjv = escape(document.bhjvmd.getJavaVendor()); var rs = bhjv=+bhjv; if (document.cookie.indexOf(bhCookieSess=1) != -1) { document.cookie = bhResults=+rs+; path=/; document.cookie = bhPrevResults=+rs+; path=/; if (document.cookie.indexOf(bhResults) != -1) self.location.replace(/html/redirect.asp?bhcp=1); else self.location.replace(/html/redirect.asp?+rs+bhqs=1); } else self.location.replace(/html/redirect.asp?+rs+bhqs=1); } // -- /script /headtitle/title noscriptbody onLoad=bhawkTest();/noscript scriptdocument.write('body onLoad=bhawkTest();');/script script language=JavaScript !-- if (navigator.javaEnabled()) document.write('applet code=JVMDetector.class name=bhjvmd width=1 height=1param name=legal value=This is copyrighted software and provided under license cyScape, Inc. (www.cyscape.com). All rights reserved./applet'); // -- /script /body/html Does anybody know someone who works for these geniuses? There's no reason this crap shouldn't work no matter what browser/platform you happen to be on, isn't this the entire point of Java? r. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
Re: [SLUG] Banks colluding with Microsoft ?
On Tue, 26 Oct 2004, Howard Lowndes wrote: On Tue, 2004-10-26 at 14:50, Rod Butcher wrote: Sluggers, I started off having to spoof my Firefox browser as Moz 1 Win98 to do St George internet banking. That stopped working so I used IE 5 Mac. That now crashes and I have to use IE6 Win XP. There is as pattern here .. and in the US i believe it would warrant antitrust / racketeering / cartel / wirefraud investigation. Anybody got any real facts on this ? Ockams Razor (paraphrased): Never ascribe to conspiracy what can be easily attributed to stupidity. occam's razor if there are many possible answers to a question, the simplest one is usually correct... Then again.. that's probably the same thing. They have XP on their desktop, so obviously they think everyone else must have to. Much more likely than a conspiracy. OTOH, it suits M$ to be non-standard, so there isn't much likelihood that they will discourage this behaviour. This is absolutely NO reason not to actively hassle the banks to conform to reasonable standards. For what it is worth, when i hassled Westpac, they actually personally phoned and mailed me with an apology and a promise that they are actually working on making their website fully standards compliant. So far nothing has changed, but I live in hope. If you don't hassle them, they will keep taking the line of least resistance. -- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html