Re: Software raid 5 performance tuning?
hi ya marc On Sat, 15 Oct 2005, Marc Dreher wrote: after a recent (data) lossy HD crash I decided to build a software raid 5 (3 disks)with Debian Sarge (2.6.13 Kernel) to keep that from happening again :-) 3 disks in raid5 means 1/3 of the total disk space is not usable Works pretty well but I am not really satsfied with the write performance (via samba). writing data into raid5 will be slow, since it's writing parts of the same data n-times to each disks reading should be faster ... but not always When uploading files via samba I get around 6 MB per second. Uploading to a non-raid disk works at full 100Mbit network throughput (10 MB p/s). presumably the target disks getting the uploaded files is on the same pc ??? and the obsvervation you're seeing is normal/typical I have one ATA disk as master on each of the two onboard IDE channels and the third SATA disk on one of my two SATA ports. odd combo but good to have only 1 disk on each ide cable I used a chunk size of 64KB with no strides. more parameters to play with So before having to experiment with all kind of settings I hope that somebody does have any tips or tuning hints for me? increasing write performance is not ez - you probably can change your nfs read/write transfer sizes to get some improvements - you probably can get faster transfers if you compress it before sending it tar zcf - stuff | ( cd /mnt/samba ; tar zxfp - ) I think with current hard- and software I should at least be able to get enough write performance to fill up my 100MBit network while uploading :-) 8bits * 10MByte/sec -- 80Mbps .. that's pretty good anything say aroun 70% - 90% of the rated networkwork thruput is good you will need to go to GigE network 8bits * 100MB/sec -- 800Mbps ... good enough for gigE entwork and the disks are rated ( by marketing folks ) at 100MB/sec (120MB/sec) or faster and your new bottleneck will be the disk rotations c ya alvin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Software raid 5 performance tuning?
Hello, after a recent (data) lossy HD crash I decided to build a software raid 5 (3 disks)with Debian Sarge (2.6.13 Kernel) to keep that from happening again :-) Works pretty well but I am not really satsfied with the write performance (via samba). When uploading files via samba I get around 6 MB per second. Uploading to a non-raid disk works at full 100Mbit network throughput (10 MB p/s). The disks are current (S)ATA Maxtor 300 Gig, CPU is AMD 64 3000, 1 Gig RAM. I have one ATA disk as master on each of the two onboard IDE channels and the third SATA disk on one of my two SATA ports. I used a chunk size of 64KB with no strides. I already did some googling and found various threads on raid performance but most of them where a few years old and did not really yield the information I was looking for. So before having to experiment with all kind of settings I hope that somebody does have any tips or tuning hints for me? I think with current hard- and software I should at least be able to get enough write performance to fill up my 100MBit network while uploading :-) Thanks best regards Marc -- Lust, ein paar Euro nebenbei zu verdienen? Ohne Kosten, ohne Risiko! Satte Provisionen für GMX Partner: http://www.gmx.net/de/go/partner -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
X performance tuning(lots of X apps at the same time)
hi I wanted to know if someone could provide me with some insight on how i could go about tuning X to run a large number of apps at once(mostly terminals). Back when running debian potato under Xfree 3.3.6 X could do this pretty easily, i would have maybe 75-100 terminals/applications open at once without much trouble. since i've upgraded to xfree 4 (woody), X seems to tap out when my process table gets to around 230 processes. that is it becomes difficult to launch new programs, and old programs(such as afterstep dock apps) seem to exit randomly. on potato i could get to 290-300 prcoesses pretty easily(same hardware) maybe a config file option? or a kernel paramter? X4 is in general less memory intensive then 3.3.6, I have to restart it about once a month when the memory usage gets to around 150-200MB. 3.3.6 I had to restart about once every 2-3 weeks with memory usage 200MB+ my process table is currentl at 229-230 processes, X memory usage is only at 70MB ..I have 161 processes running under my uid(all of which are in X), the rest are system processes, approx 260 unix sockets open on the system. I am using the same kernel as I was under potato(same compiled kernel) this has been going on ever since i upgraded around last july/august 2001. machine is p3-733 512MB, 640MB swap, matrox g400, afterstep 1.6(recompiled from potato's sources) i suppose i could just learn to close my terminals when i'm not using them anymorebut would rather tweak the system to run more! thanks nate -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: X performance tuning(lots of X apps at the same time)
quote who=nate hi I wanted to know if someone could provide me with some insight on how i could go about tuning X to run a large number of apps at once(mostly terminals). Back when running debian potato under Xfree 3.3.6 X could do this pretty easily, i would have maybe 75-100 terminals/applications open at once without much trouble. i think i found part of the problem: Jun 25 11:08:49 fury kernel: VFS: file-max limit 4096 reached i will see if i can tweak that and load a bunch of apps ... nate -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Performance tuning?
I've recently installed woody on an older system: 6x86 P150, 112MB RAM, Matrox Millennium I 4MB PCI, Buslogic multimaster PCI scsi (all scsi, no ide). This system is far from modern but I really remember it feeling a lot faster when it was my primary workstation (running win95.) I've installed XFree 4.1 and Gnome 1.2. I'm new to the the recent graphical desktop developments (the last linux that I used was in the 4.x redhat days) but it seems like everything takes forever. Clicking a tab in the Gnome preferences widget or mozilla preferences results in a 1 or 2 second delay before anything changes on the screen. Repainting the default desktop wallpaper seems to go pixel-by-pixel at times. Mozilla (0.9.9) averages about 5 to 10 seconds to render modest pages (like the debian.org home page.) This is not a function of bandwidth as I get 180 KB/s downloading packages. As an extreme example, hitting reload on the default freshmeat homepage takes mozilla ~26 seconds (more like 30+ seconds when not cached), compared to ~3 seconds for mozilla under W2k on my main workstation which uses the same internet connection. Now certainly, that machine is more modern but it's just an 800 MHz athlon. I can live with the fact that rendering large pages is cpu-bound, but back in the day this video card had very respectable 2D performance and I'm certainly not seeing any of that currently. My preferred desktop is 1152x864x24 but I'm currently at 1024x768x16 to see if it's any faster, and it's not. So my question is, what should I be looking at? Xfree is using its accelerated mga driver (I think), but how do I check to make sure it's fully tweaked? Is there anything I should check as far as bus/cpu/ram bottlenecks? I'm running 2.4.16 which I compiled for this machine. I'll doublecheck the BIOS chipset timings (i430HX) but I don't think they've changed in a long time. Disk is not a problem because the swapfile is hardly being used and the scsi subsystem is respectable. Any advice? Thanks, Brian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Performance tuning?
Brian Dessent wrote: I've recently installed woody on an older system: 6x86 P150, 112MB RAM, Matrox Millennium I 4MB PCI, Buslogic multimaster PCI scsi (all scsi, no ide). This system is far from modern but I really remember it feeling a lot faster when it was my primary workstation (running win95.) I've installed XFree 4.1 and Gnome 1.2. I'm new to the the recent graphical desktop developments (the last linux that I used was in the 4.x redhat days) but it seems like everything takes forever. I have moved to a different Window Manager on the slower systems/faster response demands because I didn't want to wait for anything. WindowMaker is available under Debian and, while it's not a Desktop like KDE/Gnome, it does do a very nice (and fast) job... -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Performance tuning?
Is this an 6x86 P150 or an 800MHz Athlon? Or is the win2k box the 800MHz Athlon? Some musings: - how much RAM on the win2k box? - On my 1GHz Athlon, I see X spiking up to 60% CPU at times. (Yes, I run X4.1, Gnome 1.2 and mozilla 1.0rc3. - Do you run gnome 1.2 or 1.4? - mozilla is pretty slow to load on my win2k box that only has 128MB RAM. - On such an old machine, best to use a light weight window manager instead of a big, fat windowing environment (and gnome is much slimmer than KDE!) Try fvwm2, blackbox, or xfce. On Sat, 2002-06-08 at 17:23, Brian Dessent wrote: I've recently installed woody on an older system: 6x86 P150, 112MB RAM, Matrox Millennium I 4MB PCI, Buslogic multimaster PCI scsi (all scsi, no ide). This system is far from modern but I really remember it feeling a lot faster when it was my primary workstation (running win95.) I've installed XFree 4.1 and Gnome 1.2. I'm new to the the recent graphical desktop developments (the last linux that I used was in the 4.x redhat days) but it seems like everything takes forever. Clicking a tab in the Gnome preferences widget or mozilla preferences results in a 1 or 2 second delay before anything changes on the screen. Repainting the default desktop wallpaper seems to go pixel-by-pixel at times. Mozilla (0.9.9) averages about 5 to 10 seconds to render modest pages (like the debian.org home page.) This is not a function of bandwidth as I get 180 KB/s downloading packages. As an extreme example, hitting reload on the default freshmeat homepage takes mozilla ~26 seconds (more like 30+ seconds when not cached), compared to ~3 seconds for mozilla under W2k on my main workstation which uses the same internet connection. Now certainly, that machine is more modern but it's just an 800 MHz athlon. I can live with the fact that rendering large pages is cpu-bound, but back in the day this video card had very respectable 2D performance and I'm certainly not seeing any of that currently. My preferred desktop is 1152x864x24 but I'm currently at 1024x768x16 to see if it's any faster, and it's not. So my question is, what should I be looking at? Xfree is using its accelerated mga driver (I think), but how do I check to make sure it's fully tweaked? Is there anything I should check as far as bus/cpu/ram bottlenecks? I'm running 2.4.16 which I compiled for this machine. I'll doublecheck the BIOS chipset timings (i430HX) but I don't think they've changed in a long time. Disk is not a problem because the swapfile is hardly being used and the scsi subsystem is respectable. -- +-+ | Ron Johnson, Jr.Home: [EMAIL PROTECTED] | | Jefferson, LA USA http://ronandheather.dhs.org:81 | | | | I have created a government of whirled peas...| | Maharishi Mahesh Yogi, 12-May-2002, | ! CNN, Larry King Live | +-+ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Performance tuning?
Ron Johnson wrote: Is this an 6x86 P150 or an 800MHz Athlon? Or is the win2k box the 800MHz Athlon? Some musings: - how much RAM on the win2k box? - On my 1GHz Athlon, I see X spiking up to 60% CPU at times. (Yes, I run X4.1, Gnome 1.2 and mozilla 1.0rc3. - Do you run gnome 1.2 or 1.4? - mozilla is pretty slow to load on my win2k box that only has 128MB RAM. - On such an old machine, best to use a light weight window manager instead of a big, fat windowing environment (and gnome is much slimmer than KDE!) Try fvwm2, blackbox, or xfce. The specs of the Win2K machine (Athlon 800 Mhz, 256 MB) are irrelevat really, I was just using it compare the render speeds of Mozilla on the two machines. Debian is installed on the 6x86 (P150, 112MB.) I will try some slimmer window managers. I know these programs need a lot of RAM but with Gnome+Sawfish+Mozilla loaded, I still have 15-20MB free RAM (accoring to 'top') so the memory footprint is not really the issue. The problem is that simple things like moving/resizing a window, drawing a menu, or repainting the background can be very slow. I was hoping that somewhere I didn't have a PCI bus or video acceration option set right. Does anyone know how much the FPU is used by these tasks? The 6x86 fpu sucks balls, but I can't really see it being used for web surfing, etc. Final question: does anyone know anything about the i430HX and its cache and tag ram? I have a 256KB pipeline-burst L2 cache module installed. I seem to recall that the stock board only included enough static tag ram to cache the first 64MB. I bought and installed a 32k x 8 x 15ns SRAM chip and installed it in the empty socket on the board. The bios has an option to set the cacheable ram to 64MB or 512MB. I cannot boot with it set to 512MB, I suspect that perhaps the 512KB L2 cache module might be required (they stopped making these many years ago I'm sure.) This all happened a long time ago so I can't really remember the details. Anyway, regarless of this cache business, w95 felt a whole lot faster than woody, so I'm looking for stuff that I haven't turned on or optimized correctly. I include below the dmesg kernel startup, in case anyone spots anything out of the ordinary. Brian (and yes, I meant Gnome 1.4 not 1.2) Linux version 2.4.16 ([EMAIL PROTECTED]) (gcc version 2.95.4 20011002 (Debian prerelease)) #3 Sat Jun 8 01:03:54 PDT 2002 BIOS-provided physical RAM map: BIOS-e820: - 0009fc00 (usable) BIOS-e820: 0009fc00 - 000a (reserved) BIOS-e820: 000f - 0010 (reserved) BIOS-e820: 0010 - 00f0 (usable) BIOS-e820: 00f0 - 0100 (reserved) BIOS-e820: 0100 - 0700 (usable) BIOS-e820: fec0 - fec01000 (reserved) BIOS-e820: fee0 - fee01000 (reserved) BIOS-e820: - 0001 (reserved) On node 0 totalpages: 28672 zone(0): 4096 pages. zone(1): 24576 pages. zone(2): 0 pages. Kernel command line: auto BOOT_IMAGE=Linux ro root=805 Initializing CPU#0 Console: colour VGA+ 132x60 Calibrating delay loop... 119.60 BogoMIPS Memory: 109576k/114688k available (1048k kernel code, 3700k reserved, 311k data, 200k init, 0k highmem) Checking if this processor honours the WP bit even in supervisor mode... Ok. Dentry-cache hash table entries: 16384 (order: 5, 131072 bytes) Inode-cache hash table entries: 8192 (order: 4, 65536 bytes) Mount-cache hash table entries: 2048 (order: 2, 16384 bytes) Buffer-cache hash table entries: 4096 (order: 2, 16384 bytes) Page-cache hash table entries: 32768 (order: 5, 131072 bytes) Enabling CPUID on Cyrix processor. CPU: Before vendor init, caps: 0001 , vendor = 1 CPU: After vendor init, caps: 0001 0004 CPU: After generic, caps: 0001 0004 CPU: Common caps: 0001 0004 CPU: Cyrix 6x86 2x Core/Bus Clock stepping 05 Checking 'hlt' instruction... OK. POSIX conformance testing by UNIFIX mtrr: v1.40 (20010327) Richard Gooch ([EMAIL PROTECTED]) mtrr: detected mtrr type: Cyrix ARR PCI: PCI BIOS revision 2.10 entry at 0xfb2b0, last bus=0 PCI: Probing PCI hardware Limiting direct PCI/PCI transfers. Activating ISA DMA hang workarounds. Linux NET4.0 for Linux 2.4 Based upon Swansea University Computer Society NET3.039 Starting kswapd Journalled Block Device driver loaded pty: 256 Unix98 ptys configured Real Time Clock Driver v1.10e block: 128 slots per queue, batch=32 Floppy drive(s): fd0 is 1.44M FDC 0 is a post-1991 82077 Linux Tulip driver version 0.9.15-pre9 (Nov 6, 2001) eth0: Lite-On PNIC-II rev 37 at 0x6100, 00:C0:F0:77:94:E6, IRQ 10. SCSI subsystem driver Revision: 1.00 scsi: * BusLogic SCSI Driver Version 2.1.15 of 17 August 1998 * scsi: Copyright 1995-1998 by Leonard N. Zubkoff [EMAIL PROTECTED] scsi0: Configuring BusLogic Model
XFree86 performance tuning
-BEGIN PGP SIGNED MESSAGE- Hey all. I was recently logged in to a DEC Unix machine in my school's lab, and I noticed that the X server was run with a negative nice value. This got me wondering what I might be able to do to get better performance out of X. I seem to recall seeing a HOWTO or something along those lines addressing this topic, but I haven't been able find it recently. Does anybody have any pointers to references that I might be able to check out for info on relatively advanced X performance tuning? I've been running X for several years now, so I'm not at all worried about tweaking the XF86Config file or using xvidtune or whatever. TIA, noah PGP public key available at http://lynx.dac.neu.edu/home/httpd/n/nmeyerha/mail.html or by 'finger -l [EMAIL PROTECTED]' -BEGIN PGP SIGNATURE- Version: 2.6.2 iQCVAwUBNm60XIdCcpBjGWoFAQG/GAP/eJILLQMqxmGtcvVXYLPukyrywo1QFhyz QWDGAR0K1q/0BBK+psDr4M/iONV33LF/RxHXiyOFHxVresYXIo51AlXZ0MGBzbsR QVIfBk5guFRxenZD1IXAVZYybdjQDhKYpBR58aDFiqxVlvUprj6WrcY6rK9PMhGt sVGIzdtoNW0= =5mgP -END PGP SIGNATURE-
Sendmail performance tuning question
I'm running into some errors trying to use some of the more advanced performance tuning features of sendmail 8.8.5. If you're responsible for a debian system running sendmail and you're using any of these features: Single Threaded Delivery - the most important one to me Host Status Directory Deferred delivery mode please contact me. Thanks! Pete -- Peter J. Templin, Jr. Client Services Analyst Computer Communication Services tel: (717) 524-1590 Bucknell University [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .