Re: zpool - low speed write
В сообщении от 5 августа 2010 13:35:04 вы написали: Write performance here is abysmal, agreed. This is very odd. I hate to say this, but can you remove ahci.ko (ahci_load=yes) from your loader.conf and reboot? You may need to change filesystem names around in /etc/fstab for your OS disk (assuming it's on ada0), but for ZFS it should just magically find the disks on adXX. If you could also provide pciconf -lvc output that would be helpful. Thanks. dd if=/dev/zero of=/tank/test.zero bs=3M count=1000 1000+0 records in 1000+0 records out 3145728000 bytes transferred in 485.431690 secs (6480269 bytes/sec) zpool iostat -v 10 10 capacity operationsbandwidth pool used avail read write read write -- - - - - - - tank1,91T 829G 0 64 49,7K 4,73M ad12 598G 333G 0 22 19,9K 1,71M ad8633G 295G 0 21 13,8K 1,65M ad10 727G 201G 0 19 15,9K 1,36M -- - - - - - - capacity operationsbandwidth pool used avail read write read write -- - - - - - - tank1,91T 829G 0116 0 5,91M ad12 598G 333G 0 39 0 2,20M ad8633G 295G 0 39 0 2,02M ad10 727G 201G 0 37 0 1,68M -- - - - - - - capacity operationsbandwidth pool used avail read write read write -- - - - - - - tank1,91T 829G 0140 0 9,09M ad12 598G 333G 0 51 0 3,69M ad8633G 295G 0 45 0 2,94M ad10 727G 201G 0 43 0 2,46M -- - - - - - - capacity operationsbandwidth pool used avail read write read write -- - - - - - - tank1,91T 829G 0133 0 7,66M ad12 598G 333G 0 46 0 2,84M ad8633G 295G 0 44 0 2,59M ad10 727G 201G 0 43 0 2,23M -- - - - - - - capacity operationsbandwidth pool used avail read write read write -- - - - - - - tank1,91T 829G 0133 6,39K 5,84M ad12 598G 333G 0 47 6,39K 2,34M ad8633G 295G 0 43 0 1,83M ad10 727G 201G 0 42 0 1,67M -- - - - - - - capacity operationsbandwidth pool used avail read write read write -- - - - - - - tank1,91T 829G 0113 0 5,54M ad12 598G 333G 0 39 0 1,97M ad8633G 295G 0 37 0 1,98M ad10 727G 201G 0 35 0 1,59M -- - - - - - - capacity operationsbandwidth pool used avail read write read write -- - - - - - - tank1,91T 829G 0152 0 10,1M ad12 598G 333G 0 52 0 3,41M ad8633G 295G 0 52 0 3,65M ad10 727G 201G 0 47 0 3,06M -- - - - - - - capacity operationsbandwidth pool used avail read write read write -- - - - - - - tank1,91T 828G 0116 0 5,61M ad12 598G 333G 0 41 0 2,16M ad8633G 295G 0 40 0 1,95M ad10 727G 201G 0 34 0 1,50M -- - - - - - - capacity operationsbandwidth pool used avail read write read write -- - - - - - - tank1,91T 828G 0176 0 11,1M ad12 598G 333G 0 60 0 3,78M ad8634G 294G 0 60 0 3,95M ad10 727G 201G 0 55 0 3,35M -- - - - - - - capacity operationsbandwidth pool used avail read write read write -- - - - - - - tank1,91T 828G 0112 0 7,55M ad12 598G 333G 0 39 0 2,73M ad8634G 294G 0 39 0 2,66M ad10 727G 201G 0 33 0 2,15M -- - - - - - - pciconf -lvc hos...@pci0:0:0:0: class=0x06 card=0x50001458 chip=0x2e208086 rev=0x02 hdr=0x00 vendor
Re: zpool - low speed write
On Thu, Aug 05, 2010 at 02:09:57PM +0800, Alex V. Petrov wrote: В сообщении от 5 августа 2010 13:35:04 вы написали: Write performance here is abysmal, agreed. This is very odd. I hate to say this, but can you remove ahci.ko (ahci_load=yes) from your loader.conf and reboot? You may need to change filesystem names around in /etc/fstab for your OS disk (assuming it's on ada0), but for ZFS it should just magically find the disks on adXX. If you could also provide pciconf -lvc output that would be helpful. Thanks. dd if=/dev/zero of=/tank/test.zero bs=3M count=1000 1000+0 records in 1000+0 records out 3145728000 bytes transferred in 485.431690 secs (6480269 bytes/sec) Can you please remove use of the zpool entirely (e.g. zpool destroy tank) and do a write test to each disk itself? E.g.: dd if=/dev/zero of=/dev/ad8 bs=64k count=100 dd if=/dev/zero of=/dev/ad10 bs=64k count=100 dd if=/dev/zero of=/dev/ad12 bs=64k count=100 I don't recommend using large block sizes (e.g. bs=1M, bs=3M). If all of the above dds show good/decent throughput, then there's something strange going on with ZFS. If this is the case, I would recommend filing a PR and posting to freebsd-fs about the problem, pointing folks to this thread. If all of the dds show bad throughput, then could you please do the following: - Provide vmstat -i output - Install ports/sysutils/smartmontools and run smartctl -a /dev/ad8, smartctl -a /dev/ad10, and smartctl -a /dev/ad12 If only one of the dds shows bad throughput, then please: - Install ports/sysutils/smartmontools and run smartctl -a /dev/XXX, where XXX is the disk which has bad throughput - Try making a ZFS pool with all 3 disks, but then do zpool offline tank XXX and then re-attempt the following dd: dd if=/dev/zero of=/tank/test.zero bs=64k count=100 And see what throughput looks like. Thanks. -- | Jeremy Chadwick j...@parodius.com | | Parodius Networking http://www.parodius.com/ | | UNIX Systems Administrator Mountain View, CA, USA | | Making life hard for others since 1977. PGP: 4BD6C0CB | ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org
Re: zpool - low speed write
Can you please remove use of the zpool entirely (e.g. zpool destroy tank) and do a write test to each disk itself? E.g.: dd if=/dev/zero of=/dev/ad8 bs=64k count=100 dd if=/dev/zero of=/dev/ad10 bs=64k count=100 dd if=/dev/zero of=/dev/ad12 bs=64k count=100 I don't have free space for move my data from zpool (1,91T) I don't recommend using large block sizes (e.g. bs=1M, bs=3M). dd if=/dev/zero of=/tank/test.zero bs=64k count=1 1+0 records in 1+0 records out 65536 bytes transferred in 50.294832 secs (13030365 bytes/sec) If all of the above dds show good/decent throughput, then there's something strange going on with ZFS. If this is the case, I would recommend filing a PR and posting to freebsd-fs about the problem, pointing folks to this thread. If all of the dds show bad throughput, then could you please do the following: - Provide vmstat -i output vmstat -i output interrupt total rate irq1: atkbd02368 0 irq6: fdc017 0 irq16: vgapci0 ath+ 1728264100 irq18: uhci2 ehci0* 2183829127 irq19: uhci4+ 427434 24 irq21: uhci1 42295 2 irq23: uhci3 ehci1 18154 1 cpu0: timer 34317326 1997 irq256: hdac01561005 90 irq257: re0 2458465143 cpu1: timer 34316042 1997 cpu3: timer 34316081 1997 cpu2: timer 34316130 1997 Total 145687410 8482 - Install ports/sysutils/smartmontools and run smartctl -a /dev/ad8, smartctl -a /dev/ad10, and smartctl -a /dev/ad12 In the first message I wrote that smatmontools installed smartd daily output: Checking health of /dev/ada2: OK Checking health of /dev/ada3: OK Checking health of /dev/ada4: OK In the logs there are no any error messages that the controller and drives smartctl -a /dev/ad8 smartctl 5.39.1 2010-01-28 r3054 [FreeBSD 8.1-STABLE amd64] (local build) Copyright (C) 2002-10 by Bruce Allen, http://smartmontools.sourceforge.net === START OF INFORMATION SECTION === Model Family: Western Digital Caviar Green family Device Model: WDC WD10EADS-00M2B0 Serial Number:WD-WCAV51709425 Firmware Version: 01.00A01 User Capacity:1 000 204 886 016 bytes Device is:In smartctl database [for details use: -P show] ATA Version is: 8 ATA Standard is: Exact ATA specification draft version not indicated Local Time is:Thu Aug 5 18:42:22 2010 KRAST SMART support is: Available - device has SMART capability. SMART support is: Enabled === START OF READ SMART DATA SECTION === SMART overall-health self-assessment test result: PASSED General SMART Values: Offline data collection status: (0x82) Offline data collection activity was completed without error. Auto Offline Data Collection: Enabled. Self-test execution status: ( 25) The self-test routine was aborted by the host. Total time to complete Offline data collection: (20400) seconds. Offline data collection capabilities:(0x7b) SMART execute Offline immediate. Auto Offline data collection on/off support. Suspend Offline collection upon new command. Offline surface scan supported. Self-test supported. Conveyance Self-test supported. Selective Self-test supported. SMART capabilities:(0x0003) Saves SMART data before entering power-saving mode. Supports SMART auto save timer. Error logging capability:(0x01) Error logging supported. General Purpose Logging supported. Short self-test routine recommended polling time:( 2) minutes. Extended self-test routine recommended polling time:( 235) minutes. Conveyance self-test routine recommended polling time:( 5) minutes. SCT capabilities: (0x303f) SCT Status supported. SCT Feature Control supported. SCT Data Table supported. SMART Attributes Data Structure revision number: 16 Vendor Specific SMART Attributes with Thresholds: ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE 1 Raw_Read_Error_Rate 0x002f 200 200 051Pre-fail Always - 0 3
Re: Kernel symbol file alternate location
Quoth Daniel O'Connor on Thursday, 05 August 2010: Is it possible to tell installkernel to put the symbol files elsewhere? IMO somewhere in /usr would be good - you don't need them to boot (or crashdump, etc) and they increase the size of the kernel by a factor of 5(!), eg.. [cain 14:07] /tmp du -sh kernel 258Mkernel [cain 14:07] /tmp rm -f kernel/*.symbols [cain 14:07] /tmp du -sh kernel 50Mkernel One downside is that you need to keep them in sync which is a bit of a drag, however I think most people use installworld for upgrades/builds so it would generally not be an issue. I think this is the main reason / has had to grow - the actual kernel is relatively small so even a 256Mb / could hold several, but with the symbol files it is not possible. I don't have patches to do this though :) -- Daniel O'Connor software and network engineer for Genesis Software - http://www.gsoft.com.au The nice thing about standards is that there are so many of them to choose from. -- Andrew Tanenbaum GPG Fingerprint - 5596 B766 97C0 0E94 4347 295E E593 DC20 7B3F CE8C +1 to this idea. I'm currently building without symbols so I don't have to clean them out of kernel.old every time I do an installkernel. -- Sterling (Chip) Camden| sterl...@camdensoftware.com | 2048D/3A978E4F http://camdensoftware.com | http://chipstips.com| http://chipsquips.com pgpFMRpmWLk0h.pgp Description: PGP signature
Re: zpool - low speed write
On Wed, Aug 4, 2010 at 9:47 PM, Alex V. Petrov alexvpet...@gmail.com wrote: ... vfs.zfs.cache_flush_disable=1 vfs.zfs.zil_disable=1 I question both of these settings, especially the latter. Please remove them both and re-test your write performance. I removed all settings of zfs. Now it default. ZFS would throttle writes if it thinks that not enough memory is available. Did you by any chance tinker with VM parameters, too? Could you post output of following commands? sysctl vm |grep kmem sysctl vfs.zfs sysctl kstat.zfs (before and after after you do some of your write speed tests) --Artem ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org
Re: Kernel symbol file alternate location
Daniel O'Connor wrote: Is it possible to tell installkernel to put the symbol files elsewhere? I wondered the same. IMO somewhere in /usr would be good - you don't need them to boot (or crashdump, etc) and they increase the size of the kernel by a factor of 5(!), eg.. [cain 14:07] /tmp du -sh kernel 258Mkernel [cain 14:07] /tmp rm -f kernel/*.symbols [cain 14:07] /tmp du -sh kernel 50Mkernel One downside is that you need to keep them in sync which is a bit of a drag, however I think most people use installworld for upgrades/builds so it would generally not be an issue. I think this is the main reason / has had to grow - the actual kernel is relatively small so even a 256Mb / could hold several, but with the symbol files it is not possible. I think a very simple solution would be to install the symbol files elsewhere (probably configurable via make.conf), and install symlinks in the kernel directory. If you do this, tools using the symbol files won't have to be changed. This would probably be a fairly trivial change to the install- kernel target, I guess. I don't have patches, though. Best regards Oliver -- Oliver Fromme, secnetix GmbH Co. KG, Marktplatz 29, 85567 Grafing b. M. Handelsregister: Registergericht Muenchen, HRA 74606, Geschäftsfuehrung: secnetix Verwaltungsgesellsch. mbH, Handelsregister: Registergericht Mün- chen, HRB 125758, Geschäftsführer: Maik Bachmann, Olaf Erb, Ralf Gebhart FreeBSD-Dienstleistungen, -Produkte und mehr: http://www.secnetix.de/bsd File names are infinite in length, where infinity is set to 255 characters. -- Peter Collinson, The Unix File System ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org
Re: 8.1 stable ar9285 ath0 problem
Hi! Can you show me your config files: rc.conf, /boot/loader.conf, wpa_supplicant.conf. I just want to see just to make sure I'm not doing it wrong. I use the stuff in a different way (mostly manually). For example: Step 1) ifconfig wlan0 up Wait a little and check: ifconfig wlan0 list scan If there's no SSID listed, wait until one comes up. Step 2) wpa_supplicant -d -d -i wlan0 -c /etc/wpa-psk.conf with /etc/wpa-psk.conf: --- ctrl_interface=/var/run/wpa_supplicant ctrl_interface_group=wheel # home network; allow all valid ciphers network={ ssid=correct-ssid key_mgmt=WPA-PSK psk=correct-psk } --- Step 3) dhclient wlan0 Now one should be able to connect to the wlan. I'm using my card on a network that uses WPA-PSK. My card can scan for APs and associate with the chosen AP, but 'dhclient wlan0' will not get an IP address! How is that possible. First wpa_supp*, then dhclient. wpa* provides the crypto so that dhclient gets the DHCP packets. And when I do 'ifconfig' it shows that the card is associated but the ip address part reads 'inet: 0.0.0.0'. Well, that's a start. -- p...@opsec.eu+49 171 310137210 years to go ! ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org
Please join us for Bugathon #7, this weekend (6th-9th August)
(bcc'd to freebsd-stable@, please reply to freebsd-current@) Hi all, Apologies for the late notice, but the bug team will be holding a bugathon this weekend. We'll be starting on Friday 6th, and continuing through Monday 9th, and the aim is to put a real effort into getting patches from PRs into a committable state, and get them committed. Some PRs will be invalid because either the bug no longer exists or was fixed without the PR being closed, identifying these is also of great benefit. The basic plan is to get as many volunteers and committers into the same place (#freebsd-bugbusters on EFnet), and work through some PRs. For this particular bugathon, we're planning on focusing on the ~1600 PRs which contain patches. Committers can help by being in the channel and available to review patches and commit them, but these events benefit especially from volunteers who are not committers: the more people who are able to look at PRs, evalaute the patches, and assess whether the patches are correct and/or the best solution to the problem, the better. If you've never been more than a user of FreeBSD then this is a great way to start to get involved - many of the bugs in the database are relatively simple to fix, and are just waiting for somebody with enough time to sit down and actually take a close look at the bugs. If you're able to evaluate patches and actually justify why the patch included is the correct solution then that is a huge help, too! So, please join us in #freebsd-bugbusters if you are free at any point over the weekend. I'll be in the channel pretty much all the time during the day (GMT) Friday - Monday, and other bugbusters/bugmeisters will be around over those four days too. Everybody is welcome to join us, the more eyes the better. We should hhopefully ave quite a few committers in the channel too, so there should be plenty of expertise available to review and commit the patches that are in a committable state. There are several wiki pages available for people who are interested in joining in, especially: http://wiki.freebsd.org/Bugathons/2010August http://wiki.freebsd.org/Bugathons/PRsWithPatches http://wiki.freebsd.org/BugBusting/Resources Many thanks, Gavin ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org
FreeBSD 7.3p1 repeatable crash when running on ramdisk
I work with a small number of FreeBSD 7.3p1 amd64 systems which are running off of an MFS root partition loaded via a DVD, and there have been a couple of problems with the mfsroot. I'd like to focus on one in particular and see if the assembled minds here have any insight as to what this might be. Basically if I do this: # yes /crashme the system doesn't panic, there's no warning (except for some random 'k' characters being output to the console) and the machine simply resets. I have some configuration details on the following URL: http://www.jetcafe.org/dave/freebsd/dvdconfig.html for perusal. The df for said machine looks like so: Filesystem SizeUsed Avail Capacity Mounted on /dev/md0c 270M231M 39M85%/ devfs 1.0K1.0K 0B 100%/dev /dev/acd0 606M606M 0B 100%/cd0 /dev/md1.uzip 254M225M 29M89%/usr /dev/md231M 30K 30M 0%/usr_rw above:/usr_rw284M254M 30M89%/usr /dev/da1s1d653G1.0G600G 0%/rw procfs 4.0K4.0K 0B 100%/proc I would expect that a system running off of MFS would inform one via log files or console messages when I fill up the root partition. I'd say it's definately a bug but I'm not sure if it's mine or not. :) Thanks in advance for any answers you all can provide. -- Dave Hayes - Consultant - Altadena CA, USA - d...@jetcafe.org The opinions expressed above are entirely my own A poor man shames us all. ___ freebsd-stable@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to freebsd-stable-unsubscr...@freebsd.org
Re: Kernel symbol file alternate location
On 06/08/2010, at 2:38, Oliver Fromme wrote: I think this is the main reason / has had to grow - the actual kernel is relatively small so even a 256Mb / could hold several, but with the symbol files it is not possible. I think a very simple solution would be to install the symbol files elsewhere (probably configurable via make.conf), and install symlinks in the kernel directory. If you do this, tools using the symbol files won't have to be changed. This would probably be a fairly trivial change to the install- kernel target, I guess. I don't have patches, though. Yeah, I don't think it's hard to move them, however I'm worried what it will break :) The only thing I can see that would have to change would be kgdb so it tells gdb where to find the symbols. -- Daniel O'Connor software and network engineer for Genesis Software - http://www.gsoft.com.au The nice thing about standards is that there are so many of them to choose from. -- Andrew Tanenbaum GPG Fingerprint - 5596 B766 97C0 0E94 4347 295E E593 DC20 7B3F CE8C