Re: FreeBSD NFS client/Linux NFS server issue
Hi, In this thread I have posted to freebsd-fs@ several messages describing our problem with freebsd7.1 nfs clients. As with the time new info has appeared and having this all spread in several messages might be a bit confusing, I want to summarise here what we see and know. Also I cc to freebsd-stable@ hoping to draw more attention to this problem as it looks for me very interesting and challenging :-) I have found in the Internet that other people have been observed the similar problem with FreeBSD6.2 client: http://forums.freebsd.org/showthread.php?t=1697 So, on some of our freebsd7.1 nfs clients (and it looks like we have had similar case with 6.3), which have several nfs mounts to the same CentOS 5.3 NFS server (mount options: rw,-3,-T,-s,-i,-r=32768,-w=32768,-o=noinet6), at some moment the access to one of the NFS mount gets stuck, while the access to the other mounts works ok. In all cases we have been observed so far the first gotten stuck process was php script (or two) that was (were) writing to logs file (appending). In tcpdump we see that every write to the file causes the sequence of the following rpc: ACCESS - READ - WRITE - COMMIT. And at some moment this stops after READ rpc call and successful reply. After this in tcpdump successful readdir/access/lookup/fstat calls are observed from our other utilities, which just check the presence of some files and they work ok (df also works). The php process at this state is in bo_wwait invalidating buffer cache [1]. If at this time we try accessing the share with mc then it hangs acquiring the vn_lock held by php process [2] and after this any operations with this NFS share hang (df hangs too). If instead some other process is started that writes to some other file on this share (append) then the first process unfreezes too (starting from WRITE rpc, so there is no any retransmits). With my limited knowledge of this complicated kernel subsystem I have the following hypothesis what is going on. On some of the nfs_write() it does successful ACCESS - READ rpcs but by some reason does not call WRITE to flush dirty buffer to the server (aborts somewere or may be in bdwrite() which calls bd_wakeup() and actually bd_wakeup considers that we don't have enough dirty buffers?). But it looks like on this stage the buffer appears to be unlinked from bufqueues [3] so when bufdaemon runs it does not flush the buffer. The next write() call to this file causes the process to get stuck invalidating the dirty buffer. The buffer is accessible by nfsiod via nmp structure [3] and when the next process is writing to another file, nfsiod is started and flushes this dirty buffer. [1]: Gotten stuck php process: (kgdb) bt #0 sched_switch (td=0xc839e000, newtd=Variable newtd is not available. ) at /usr/src/sys/kern/sched_ule.c:1944 #1 0xc07cabe6 in mi_switch (flags=Variable flags is not available. ) at /usr/src/sys/kern/kern_synch.c:440 #2 0xc07f42fb in sleepq_switch (wchan=Variable wchan is not available. ) at /usr/src/sys/kern/subr_sleepqueue.c:497 #3 0xc07f460c in sleepq_catch_signals (wchan=0xc90c9ee8) at /usr/src/sys/kern/subr_sleepqueue.c:417 #4 0xc07f4ebd in sleepq_wait_sig (wchan=0xc90c9ee8) at /usr/src/sys/kern/subr_sleepqueue.c:594 #5 0xc07cb047 in _sleep (ident=0xc90c9ee8, lock=0xc90c9e8c, priority=333, wmesg=0xc0b731ed bo_wwait, timo=0) at /usr/src/sys/kern/kern_synch.c:224 #6 0xc0827295 in bufobj_wwait (bo=0xc90c9ec4, slpflag=256, timeo=0) at /usr/src/sys/kern/vfs_bio.c:3870 #7 0xc0966307 in nfs_flush (vp=0xc90c9e04, waitfor=1, td=0xc839e000, commit=1) at /usr/src/sys/nfsclient/nfs_vnops.c:2989 #8 0xc09667ce in nfs_fsync (ap=0xed3c38ec) at /usr/src/sys/nfsclient/nfs_vnops.c:2725 #9 0xc0aee5d2 in VOP_FSYNC_APV (vop=0xc0c2b920, a=0xed3c38ec) at vnode_if.c:1007 #10 0xc0827864 in bufsync (bo=0xc90c9ec4, waitfor=1, td=0xc839e000) at vnode_if.h:538 #11 0xc083f354 in bufobj_invalbuf (bo=0xc90c9ec4, flags=1, td=0xc839e000, slpflag=256, slptimeo=0) at /usr/src/sys/kern/vfs_subr.c:1066 #12 0xc083f6e2 in vinvalbuf (vp=0xc90c9e04, flags=1, td=0xc839e000, slpflag=256, slptimeo=0) at /usr/src/sys/kern/vfs_subr.c:1142 #13 0xc094f216 in nfs_vinvalbuf (vp=0xc90c9e04, flags=Variable flags is not available. ) at /usr/src/sys/nfsclient/nfs_bio.c:1326 #14 0xc0951825 in nfs_write (ap=0xed3c3bc4) at /usr/src/sys/nfsclient/nfs_bio.c:918 #15 0xc0aef956 in VOP_WRITE_APV (vop=0xc0c2b920, a=0xed3c3bc4) at vnode_if.c:691 #16 0xc0850097 in vn_write (fp=0xc9969b48, uio=0xed3c3c60, active_cred=0xcb901600, flags=0, td=0xc839e000) at vnode_if.h:373 #17 0xc07f9d17 in dofilewrite (td=0xc839e000, fd=6, fp=0xc9969b48, auio=0xed3c3c60, offset=-1, flags=0) at file.h:256 #18 0xc07f9ff8 in kern_writev (td=0xc839e000, fd=6, auio=0xed3c3c60) at /usr/src/sys/kern/sys_generic.c:401 #19 0xc07fa06f in write (td=0xc839e000, uap=0xed3c3cfc) at /usr/src/sys/kern/sys_generic.c:317 #20 0xc0ad9c75 in syscall (frame=0xed3c3d38) at
Re: immense delayed write to file system (ZFS and UFS2), performance issues
On Mon, 18 Jan 2010 21:41:53 -0500 Garrett Moore garrettmo...@gmail.com wrote about Re: immense delayed write to file system (ZFS and UFS2), performance issues: GM The drives being discussed in my related thread (regarding poor GM performance) are all WD Green drives. I have used wdidle3 to set all GM of my drive timeouts to 5 minutes. I'll see what sort of difference GM this makes for performance. GM Even if it makes no difference to performance, thank you for pointing GM it out GM -- my drives have less than 2,000 hours on them and were all over GM 90,000 load cycles due to this moronic factory setting. Since changing GM the timeout, they haven't parked (which is what I would expect). Thanks for bringing up this topic here. I have drives showing up close to 80 load cycle counts here. Guess it's time for that fix... :-| cu Gerrit ___ 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: FreeBSD NFS client/Linux NFS server issue
On Tue, 19 Jan 2010 10:02:57 +0200 Mikolaj Golub wrote: I have found in the Internet that other people have been observed the similar problem with FreeBSD6.2 client: http://forums.freebsd.org/showthread.php?t=1697 Reading this through carefully it looks like the guy did not experience the problem (gotten stuck processes). He just described the behaviour of freebsd client when appending the file. -- Mikolaj Golub ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
Garrett Moore wrote: The drives being discussed in my related thread (regarding poor performance) are all WD Green drives. I have used wdidle3 to set all of my drive timeouts to 5 minutes. I'll see what sort of difference this makes for performance. Even if it makes no difference to performance, thank you for pointing it out -- my drives have less than 2,000 hours on them and were all over 90,000 load cycles due to this moronic factory setting. Since changing the timeout, they haven't parked (which is what I would expect). You're welcome. I just feel as bad for you as for everyone else who has bought these obviously Windoze optimized harddrives. Unfortunately neither wdidle3 nor an updated firmware is available or functioning on the latest models in the Green series. At least that's what I've read from other people having this issue. WD only claims they don't support Linux and they probably have never heard of FreeBSD. If anyone successfully has fixed their WD15EADS drives this way I'd be interested in hearing from you. One of my drives has 216,000 load cycles accumulated over 8 months. That's one every 2nd minute... and I was hit by the Seagate 7200.11 fiasco too. Running on Samsungs now :-) ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
2010/1/18 Morgan Wesström freebsd-questi...@pp.dyndns.biz O. Hartmann wrote: I realise a strange behaviour of several FreeBSD 8.0-STABLE/amd64 boxes. All boxes have the most recent STABLE. One box is a UP system, two others SMP boxes, one with a Q6600 4-core, another XEON with 2x 4-cores (Dell Poweredge III). Symptome: All boxes have ZFS and UFS2 filesystems. Since two weeks or so, sometimes the I/O performance drops massively when doing 'svn update', 'make world' or even 'make kernel'. It doesn't matter what memory and how many cpu the box has, it get stuck for several seconds and freezing. On the UP box, this is sometimes for 10 - 20 seconds. A very interesting phenomenon is the massively delayed file writing on ZFS filesystems I realise. Editing a file in 'vi' running on one XTerm and having in another Xterminal my shell for compiling this file, it takes sometimes up to 20 seconds to get the file updated after it has been written. It's like having an old, slow NFS connection with long cache delays. These massively delayed file transactions are not necessarely under heavy load, sometimes they occur in a relaxed situation. They seem to occur much more often on the UP box than on the SMP boxes, but this strange phenomenon also occur on the Dell Poweredge II, which has 16GB RAM and summa summarum 16 cores. This phenomenon does occur on ZFS- and UFS2 filesystems as well. It is hardly reproducable. Is there any known issue? Ragrds, Oliver The disks involved don't happen to be Western Digital Green Power disks, do they? The Intelli-Park function in these disks are wrecking havoc with I/O in Linux-land at least, causing massive stalls and iowait through the roof during the 25-30 seconds it takes for the heads to unload after parking. I have two of these disks sitting on my desk now collecting dust... /Morgan ___ freebsd-questi...@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org ZFS is copy on write, therefore to optimize the write performance it delays writes for a long as possible, upto a set maximum time. It will then flush to the disks. How long this time is depends on how much free ram you have available. Assuming processes are eating up all your ram I would imagine you are hitting the max limit. I'm not sure exactly what its set to on bsd but I know the default on opensolaris is 30s. I think this explains your delayed writes. Not sure what will cause the lock ups though. ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
Gerrit Kühn wrote: Thanks for bringing up this topic here. I have drives showing up close to 80 load cycle counts here. Guess it's time for that fix... :-| Just note that the utility is officially for WD's Raid Edition GP drives and not for the regular consumer models although some users have reported success on using it on those too. /Morgan ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
On Tue, Jan 19, 2010 at 10:28:58AM +0100, Morgan Wesström wrote: Garrett Moore wrote: The drives being discussed in my related thread (regarding poor performance) are all WD Green drives. I have used wdidle3 to set all of my drive timeouts to 5 minutes. I'll see what sort of difference this makes for performance. Even if it makes no difference to performance, thank you for pointing it out -- my drives have less than 2,000 hours on them and were all over 90,000 load cycles due to this moronic factory setting. Since changing the timeout, they haven't parked (which is what I would expect). You're welcome. I just feel as bad for you as for everyone else who has bought these obviously Windoze optimized harddrives. Unfortunately neither wdidle3 nor an updated firmware is available or functioning on the latest models in the Green series. At least that's what I've read from other people having this issue. WD only claims they don't support Linux and they probably have never heard of FreeBSD. No offence intended by this statement, but: the Green drives are specifically intended for workstations. I don't believe in the whole segregation of drive model thing, but the fact of the matter is, the Green drives are variable-RPM and have numerous firmware-level features which intend for them to be used in workstation environments -- that means not constant I/O or heavy workload. Windows has nothing to do with this. If you want a consumer-edition drive that's better tuned for server work, you should really be looking at the WD Caviar Black series or their RE/RE2 series. I have both the Green and Black drives, and I've done my share of benchmarking. Sustained transfer rates on the Black models exceed that of the Greens by almost 20-25MB/sec. Average seek times are slightly lower, and I/O concurrency is handled much better. The Black drives also have a feature called TLER[1], which can be toggled using a utility from Western Digital. Those using these drives in a RAID or ZFS array will be very interested in disabling this feature to ensure quick timeouts from the drive in the case of I/O errors. Other manufacturers have the same feature, just called something else (ex. Samsung's is called CCTL). Now, admittedly WD doesn't give this utility out any more (which is silly), and some people have reported that recent-day (circa mid-to-late 2009) Black drives refuse to let you toggle TLER. The latter claim is absurd -- I purchased 4 Black drives (2 with manufacturing dates of October 2009, and 2 with September 2009) and all of them let me toggle TLER without any problem. Keep in mind you your SATA controller has to be set to non-AHCI mode (sometimes called Enhanced mode) or Compatibility mode (e.g. IDE emulation) for the utility work. If you have qualms/concerns/issues with Western Digital drives or their mentality behind their drives, simply don't buy them. Really. People often ask me (and others) what brand of hard disk is good? and I always tell them the same thing: there's no correct answer. Everyone has their own experience with different vendors, models, etc. [1]: http://en.wikipedia.org/wiki/Time-Limited_Error_Recovery If anyone successfully has fixed their WD15EADS drives this way I'd be interested in hearing from you. One of my drives has 216,000 load cycles accumulated over 8 months. That's one every 2nd minute... and I was hit by the Seagate 7200.11 fiasco too. Running on Samsungs now :-) Aren't Samsung's drives known for firmware bugs/quirks? The documentation associated with smartmontools discusses this quite a bit. This is one reason why I stay away from them. Fujitsu is another vendor I want absolutely nothing to do with (very high failure rates in addition to bad sectors with their SCSI-3 disks at my workplace). -- | 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: immense delayed write to file system (ZFS and UFS2), performance issues
On Tue, 19 Jan 2010 01:57:36 -0800 Jeremy Chadwick free...@jdc.parodius.com wrote about Re: immense delayed write to file system (ZFS and UFS2), performance issues: JC If you want a consumer-edition drive that's better tuned for server JC work, you should really be looking at the WD Caviar Black series or JC their RE/RE2 series. That's exactly what I did. I have WD-RE2 drives here that show exactly this problem (RE2/GP)! The model number is WD1000FYPS-01ZKB0. cu Gerrit ___ 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: [PATCH] Lockmgr deadlock on STABLE_8
May you post your kernel config? sure... include GENERIC ident DEBUG options KDB options DDB options WITNESS options INVARIANT_SUPPORT options INVARIANTS That one doesnt lockup, or didnt for the few days I tried it. The other config is the same, except missing the 'WITNESS' line. The one which locked up was just standard GENERIC ever time. It locked up on 8.0 release, and has been locking every tme I have tried it since. The symptoms look very much like it is the disc subsystem which is locking up (but eventually everything else rinds to a halt too). I updated to SVN rev 202576 last night, and have just done a reboot with a new world and GENERIC kernel built from that. Will see how long that lasts - I expect it to have locked in the next 24 hours or so though. For reference, the dmesg from the machine is attached below. Thanks for taking an interest in this one! -pete. --- Copyright (c) 1992-2010 The FreeBSD Project. Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994 The Regents of the University of California. All rights reserved. FreeBSD is a registered trademark of The FreeBSD Foundation. FreeBSD 8.0-STABLE #0: Mon Jan 18 23:25:15 GMT 2010 webad...@florentine.rattatosk:/usr/obj/usr/src/sys/GENERIC amd64 Timecounter i8254 frequency 1193182 Hz quality 0 CPU: Intel(R) Xeon(R) CPU E5345 @ 2.33GHz (2333.43-MHz K8-class CPU) Origin = GenuineIntel Id = 0x6fb Stepping = 11 Features=0xbfebfbffFPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CLFLUSH,DTS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,PBE Features2=0x4e3bdSSE3,DTES64,MON,DS_CPL,VMX,EST,TM2,SSSE3,CX16,xTPR,PDCM,DCA AMD Features=0x2800SYSCALL,LM AMD Features2=0x1LAHF TSC: P-state invariant real memory = 4294967296 (4096 MB) avail memory = 4104093696 (3913 MB) ACPI APIC Table: HP ProLiant FreeBSD/SMP: Multiprocessor System Detected: 4 CPUs FreeBSD/SMP: 1 package(s) x 4 core(s) cpu0 (BSP): APIC ID: 0 cpu1 (AP): APIC ID: 1 cpu2 (AP): APIC ID: 2 cpu3 (AP): APIC ID: 3 ACPI Warning: Invalid length for Pm1aControlBlock: 32, using default 16 20090521 tbfadt-707 ioapic0 Version 2.0 irqs 0-23 on motherboard ioapic1 Version 2.0 irqs 24-47 on motherboard kbd1 at kbdmux0 acpi0: HP ProLiant on motherboard acpi0: [ITHREAD] acpi0: Power Button (fixed) Timecounter ACPI-fast frequency 3579545 Hz quality 1000 acpi_timer0: 24-bit timer at 3.579545MHz port 0x908-0x90b on acpi0 acpi_hpet0: High Precision Event Timer iomem 0xfed0-0xfed003ff on acpi0 Timecounter HPET frequency 14318180 Hz quality 900 pcib0: ACPI Host-PCI bridge on acpi0 pci0: ACPI PCI bus on pcib0 pcib1: ACPI PCI-PCI bridge at device 2.0 on pci0 ACPI Warning: \\_SB_.PCI0.PT02._PRT: Return Package has no elements (empty) 20090521 nspredef-545 pci9: ACPI PCI bus on pcib1 pcib2: ACPI PCI-PCI bridge at device 0.0 on pci9 pci10: ACPI PCI bus on pcib2 pcib3: ACPI PCI-PCI bridge at device 0.0 on pci10 pci11: ACPI PCI bus on pcib3 pcib4: PCI-PCI bridge at device 1.0 on pci10 pci14: PCI bus on pcib4 pcib5: PCI-PCI bridge at device 2.0 on pci10 pci15: PCI bus on pcib5 pcib6: ACPI PCI-PCI bridge at device 0.3 on pci9 pci16: ACPI PCI bus on pcib6 pcib7: ACPI PCI-PCI bridge at device 3.0 on pci0 pci6: ACPI PCI bus on pcib7 pcib8: PCI-PCI bridge at device 0.0 on pci6 pci7: PCI bus on pcib8 pcib9: PCI-PCI bridge at device 4.0 on pci7 pci8: PCI bus on pcib9 ciss0: HP Smart Array E200i port 0x4000-0x40ff mem 0xfde8-0xfdef,0xfde7-0xfde77fff irq 16 at device 8.0 on pci7 ciss0: PERFORMANT Transport ciss0: got 2 MSI messages] ciss0: [ITHREAD] pcib10: ACPI PCI-PCI bridge at device 4.0 on pci0 pci19: ACPI PCI bus on pcib10 pcib11: PCI-PCI bridge at device 5.0 on pci0 pci22: PCI bus on pcib11 pcib12: ACPI PCI-PCI bridge at device 6.0 on pci0 pci2: ACPI PCI bus on pcib12 pcib13: ACPI PCI-PCI bridge at device 0.0 on pci2 pci3: ACPI PCI bus on pcib13 bce0: HP NC373i Multifunction Gigabit Server Adapter (B2) mem 0xf800-0xf9ff irq 18 at device 0.0 on pci3 miibus0: MII bus on bce0 brgphy0: BCM5708C 10/100/1000baseTX PHY PHY 1 on miibus0 brgphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto bce0: Ethernet address: 00:1e:0b:5f:1f:76 bce0: [ITHREAD] bce0: ASIC (0x57081020); Rev (B2); Bus (PCI-X, 64-bit, 133MHz); B/C (1.9.6); Flags (MSI|MFW); MFW () pcib14: ACPI PCI-PCI bridge at device 7.0 on pci0 pci4: ACPI PCI bus on pcib14 pcib15: ACPI PCI-PCI bridge at device 0.0 on pci4 pci5: ACPI PCI bus on pcib15 bce1: HP NC373i Multifunction Gigabit Server Adapter (B2) mem 0xfa00-0xfbff irq 19 at device 0.0 on pci5 miibus1: MII bus on bce1 brgphy1: BCM5708C 10/100/1000baseTX PHY PHY 1 on miibus1 brgphy1: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto bce1: Ethernet address: 00:1e:0b:5f:fd:d8 bce1: [ITHREAD] bce1: ASIC (0x57081020); Rev
Re: [PATCH] Lockmgr deadlock on STABLE_8
2010/1/19 Pete French petefre...@ticketswitch.com: May you post your kernel config? sure... include GENERIC ident DEBUG options KDB options DDB options WITNESS options INVARIANT_SUPPORT options INVARIANTS Ok then, remove the debugging (WITNESS, INVARIANT*), leave in place KDB and DDB, add GDB and try at least to get a coredump when it deadlocks. Attilio -- Peace can only be achieved by understanding - A. Einstein ___ 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: [PATCH] Lockmgr deadlock on STABLE_8
Ok then, remove the debugging (WITNESS, INVARIANT*), leave in place KDB and DDB, add GDB and try at least to get a coredump when it deadlocks. OK, will do. Am building a KDB/GDB/DDB kenel now. -pete. ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
Jeremy Chadwick wrote: No offence intended by this statement, but: the Green drives are specifically intended for workstations. I don't believe in the whole segregation of drive model thing, but the fact of the matter is, the Green drives are variable-RPM and have numerous firmware-level features which intend for them to be used in workstation environments -- that means not constant I/O or heavy workload. Windows has nothing to do with this. No offence taken. You are one of the few highly regarded contributors to this list and I always appreciate the information you share. My comment about Windows was based on the fact that Windows-users don't experience this behaviour. The 8 second Intelli-Park timeout is most probably tuned to the Windows kernel flush period but I can't say for sure. But obviously the load/unload cycle isn't triggered nearly as often on that OS. I just feel frustrated that once again a manufacturer simply ignore the users who chose to run something else on their computers than the products from Redmond. In my case I have several decades worth of experience and knowledge and I have no problem grasping the various technical features and their usefullness. The problem is that I as a consumer have no way of knowing that a feature like this would be incompatible in my environment despite my experience. In my search to track this problem down I sidetracked several times in both filesystem and kernel internals until I finally stumbled over the Intelli-Park feature. It hadn't even crossed my mind that the problem could be related to a hardware feature and that bothers me immensely... but I'll deal with it eventually ;) Aren't Samsung's drives known for firmware bugs/quirks? The documentation associated with smartmontools discusses this quite a bit. This is one reason why I stay away from them. Fujitsu is another vendor I want absolutely nothing to do with (very high failure rates in addition to bad sectors with their SCSI-3 disks at my workplace). I think both you and I know by now that every harddrive manufacturer have their problems with certain models at some point. What surprises me is the total lack of understanding from the manufacturers, most recently the firmware issue of Seagates 7200.11 series where they simply try to ignore the problem. The only professional response I've seen during my career is IBM's replacement of their Deathstar drives many years ago. I accept and understand that a manufacturer now and then makes a bad product but it's how they deal with the situation when it occurs, that decides whether I will ever buy their products again. With Samsung I've completed the full circle and if I run into a problem with them and they don't take their responsibility, well, then I'm not sure what to buy next time. Maybe I should go for Maxtor again... ;-) Regards Morgan ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
On Tue, Jan 19, 2010 at 11:07:24AM +0100, Gerrit Kühn wrote: On Tue, 19 Jan 2010 01:57:36 -0800 Jeremy Chadwick free...@jdc.parodius.com wrote about Re: immense delayed write to file system (ZFS and UFS2), performance issues: JC If you want a consumer-edition drive that's better tuned for server JC work, you should really be looking at the WD Caviar Black series or JC their RE/RE2 series. That's exactly what I did. I have WD-RE2 drives here that show exactly this problem (RE2/GP)! The model number is WD1000FYPS-01ZKB0. I should have been more specific. WD makes RE-series drives which don't have GP applied to them; those are what I was referring to. Here's the current list of Western Digital SATA models out there, sans some earlier revisions. All below drives are 3.5, SATA300, and have their SATA power/signal ports placed to permit hot-swapping in drive carriers unless otherwise noted. Enterprise high-performance class === WD3000GLFS - WD VelociRatpor, 300GB, 16MB, 10Krpm, no-hotswap WD1500HLFS - WD VelociRaptor, 150GB, 16MB, 10Krpm WD3000HLFS - WD VelociRaptor, 300GB, 16MB, 10Krpm WD1500BLFS - WD VelociRaptor, 150GB, 16MB, 10Krpm, 2.5 bare WD3000BLFS - WD VelociRaptor, 300GB, 16MB, 10Krpm, 2.5 bare Enterprise business-critical class WD2502ABYS - WD RE3, 250GB, 16MB, 7200rpm WD3202ABYS - WD RE3, 320GB, 16MB, 7200rpm WD5002ABYS - WD RE3, 500GB, 16MB, 7200rpm WD1002FBYS - WD RE3, 1TB, 32MB, 7200rpm WD7502ABYS - WD RE3, 750GB, 32MB, 7200rpm WD2003FYYS - WD RE4, 2TB, 64MB, 7200rpm Enterprise energy-saving class WD5000ABPS - WD RE2-GP, 500GB, 16MB, variable rpm WD7500AYPS - WD RE2-GP, 750GB, 16MB, variable rpm WD1000FYPS - WD RE2-GP, 1TB, 16MB, variable rpm WD2002FYPS - WD RE4-GP, 2TB, 64MB, variable rpm Desktop class === WD800AAJS - WD Caviar Blue, 80GB, 8MB, 7200rpm WD1600AAJS - WD Caviar Blue, 160GB, 8MB, 7200rpm WD2500AAJS - WD Caviar Blue, 250GB, 8MB, 7200rpm WD3200AAJS - WD Caviar Blue, 320GB, 8MB, 7200rpm WD2500AAKS - WD Caviar Blue, 250GB, 16MB, 7200rpm WD3200AAKS - WD Caviar Blue, 320GB, 16MB, 7200rpm WD5000AAKS - WD Caviar Blue, 500GB, 16MB, 7200rpm WD6400AAKS - WD Caviar Blue, 640GB, 16MB, 7200rpm WD5000AACS - WD Caviar Green, 500GB, 16MB, variable rpm WD6400AACS - WD Caviar Green, 640GB, 16MB, variable rpm WD7500AACS - WD Caviar Green, 750GB, 16MB, variable rpm WD10EACS - WD Caviar Green, 1TB, 16MB, variable rpm WD5000AADS - WD Caviar Green, 500GB, 32MB, variable rpm WD6400AADS - WD Caviar Green, 640GB, 32MB, variable rpm WD7500AADS - WD Caviar Green, 750GB, 32MB, variable rpm WD10EADS - WD Caviar Green, 1TB, 32MB, variable rpm WD15EADS - WD Caviar Green, 1.5TB, 32MB, variable rpm WD20EADS - WD Caviar Green, 2TB, 32MB, variable rpm WD6400AARS - WD Caviar Green, 640GB, 64MB, variable rpm WD8000AARS - WD Caviar Green, 800GB, 64MB, variable rpm WD10EARS - WD Caviar Green, 1TB, 64MB, variable rpm WD15EARS - WD Caviar Green, 1.5TB, 64MB, variable rpm WD20EARS - WD Caviar Green, 2TB, 64MB, variable rpm WD5001AALS - WD Caviar Black, 500GB, 32MB, 7200rpm WD6401AALS - WD Caviar Black, 640GB, 32MB, 7200rpm WD7501AALS - WD Caviar Black, 750GB, 32MB, 7200rpm WD1001FALS - WD Caviar Black, 1TB, 32MB, 7200rpm WD2001FAAS - WD Caviar Black, 2TB, 64MB, 7200rpm So which drive models above are experiencing a continual increase in SMART attribute 193 (Load Cycle Count)? My guess is that some of the WD Caviar Green models, and possibly all of the RE2-GP and RE4-GP models are experiencing this problem. I say some with regards to WD Caviar Green since I have some which do not appear to exhibit the heads/actuator arm moved into the landing/park zone. I'm at work right now, but when I get home I can verify what models I've used which didn't experience this problem, as well as what the manufacturing date and F/W revisions are. I should note I don't have said Green drives in use (I use WD1001FALS drives now). -- | 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: immense delayed write to file system (ZFS and UFS2), performance issues
On Tue, 19 Jan 2010 03:24:49 -0800 Jeremy Chadwick free...@jdc.parodius.com wrote about Re: immense delayed write to file system (ZFS and UFS2), performance issues: JC JC If you want a consumer-edition drive that's better tuned for JC JC server work, you should really be looking at the WD Caviar Black JC JC series or their RE/RE2 series. JC That's exactly what I did. I have WD-RE2 drives here that show JC exactly this problem (RE2/GP)! The model number is WD1000FYPS-01ZKB0. JC I should have been more specific. WD makes RE-series drives which JC don't have GP applied to them; those are what I was referring to. Well, when I bought these drives I was not aware of this issue. Buying a drive intended for 24/7 use in RAID configurations is basically the right idea, I think. From what was written about the GP feature back then I could not anticipate such problems. I would have liked to buy the 2TB drives without GP lately, but they have lead times into April here. So I went for the GP model, which now shows the same problem as the 1TB drive... :-( JC WD1000FYPS - WD RE2-GP, 1TB, 16MB, variable rpm JC WD2002FYPS - WD RE4-GP, 2TB, 64MB, variable rpm JC So which drive models above are experiencing a continual increase in JC SMART attribute 193 (Load Cycle Count)? My guess is that some of the JC WD Caviar Green models, and possibly all of the RE2-GP and RE4-GP JC models are experiencing this problem. I can confirm that the two models above show this problem. Furthermore I can confirm that at least in my setup here this drive type works fine: WD5001ABYS I have some of the RE3 drives sitting around here and will probably try them later. Can anyone here report anything about the fixed firmware from http://support.wdc.com/product/download.asp?groupid=609sid=113lang=en? Does this remedy the problem for the 1TB RE2 drive? JC I say some with regards to WD Caviar Green since I have some which do JC not appear to exhibit the heads/actuator arm moved into the JC landing/park zone. I'm at work right now, but when I get home I can JC verify what models I've used which didn't experience this problem, as JC well as what the manufacturing date and F/W revisions are. I should JC note I don't have said Green drives in use (I use WD1001FALS drives JC now). Thanks for sharing this information. cu Gerrit ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
On Tue, Jan 19, 2010 at 09:16:41AM +0100, Gerrit K?hn wrote: Thanks for bringing up this topic here. I have drives showing up close to 80 load cycle counts here. Guess it's time for that fix... :-| Device Model: WDC WD10EACS-00ZJB0 Firmware Version: 01.01B01 Serial Number:WD-WCAS [...] 9 Power_On_Hours 17046 193 Load_Cycle_Count 1045512 The above drive is in a raidz of three. The other two drives from that batch have already failed. :( In another system: Device Model: WDC WD10EACS-00D6B0 Firmware Version: 01.01A01 Serial Number:WD-WCAU [...] 9 Power_On_Hours 13111 193 Load_Cycle_Count 7 ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
Emil Mikulic wrote: On Tue, Jan 19, 2010 at 09:16:41AM +0100, Gerrit K?hn wrote: Thanks for bringing up this topic here. I have drives showing up close to 80 load cycle counts here. Guess it's time for that fix... :-| Device Model: WDC WD10EACS-00ZJB0 Firmware Version: 01.01B01 Serial Number:WD-WCAS [...] 9 Power_On_Hours 17046 193 Load_Cycle_Count 1045512 The above drive is in a raidz of three. The other two drives from that batch have already failed. :( Did you RMA the failing drives? Did WD comment the Load_Cycle_Count? /Morgan ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
On Tue, Jan 19, 2010 at 03:24:49AM -0800, Jeremy Chadwick wrote: WD2001FAAS - WD Caviar Black, 2TB, 64MB, 7200rpm Do you mean WD2001FASS? I can't find a WD2001FAAS. Thanks, Gary ___ 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
8.0 stable if_bwi kmod not exist?
folks, There is not exist if_bwi.ko module in /boot/kernel under 8.0 Stable why? ___ 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
scp from livecd (fixit option) cd can not find /usr/bin/ssh
Hello i use the livecd to save an old pc. But when i try to use scp to copy some data to our network i get the message that /usr/bin/ssh could not be found. This is on 8.0 i386 livecd Regards, Johan Hendriks ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
On Tue, Jan 19, 2010 at 10:44:59AM -0500, Gary Palmer wrote: On Tue, Jan 19, 2010 at 03:24:49AM -0800, Jeremy Chadwick wrote: WD2001FAAS - WD Caviar Black, 2TB, 64MB, 7200rpm Do you mean WD2001FASS? I can't find a WD2001FAAS. Yup, typo -- bound to be at least one given the amount of data I typed in. :-) Good catch! -- | 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
Pack of CAM improvements
Hi. I've made a patch, that should solve set of problems of CAM ATA and CAM generally. I would like to ask for testing and feedback. What patch does: - It unifies bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices). - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several run levels. Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. Patch can be found here: http://people.freebsd.org/~mav/cam-ata.20100119.patch Feedback as always welcome. -- Alexander Motin ___ 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: ZFS performance degradation over time
I've been watching my memory usage and I have no idea what is consuming memory as 'Active'. Last night I had around 6500MB 'Active' again, 1500MB Wired, no inact, ~30MB buf, no free, and ~100MB swap used. My performance copying ZFS-ZFS was again slow (1MB/s). I tried killing rTorrent and no significant amount of memory was reclaimed - maybe 100MB. `ps aux` showed no processes using any significant amount of memory, and I was definitely nowhere near 6500MB usage. I tried running the perl oneliner again to hog a bunch of memory, and almost all of the Active memory was IMMEDIATELY marked as Free, and my performance was excellent again. I'm not sure what in userland could be causing the issue. The only things I've installed are rTorrent, lighttpd, samba, smartmontools, vim, bash, Python, Perl, and SABNZBd. There is nothing that *should* be consuming any serious amount of memory. On Mon, Jan 18, 2010 at 11:29 AM, Norbert Papke npa...@acm.org wrote: On January 17, 2010, Garrett Moore wrote: I upgraded my system to 8GB of ram to see if that would help. It hasn't made much of a difference. After having rTorrent running for a while, my performance again tanked. Around 6.5GB of memory was showing as 'Active' according to top. 6.5GB of active memory seems to imply that a user process is growing or a large number of user processes are being created. I would expect ZFS's cache to increase the size of wired memory. Sorry, I have not followed this thread closely. Are you sure that the degradation is ZFS related? Could it be caused by, for instance, a userland memory leak? What happens to active memory when you restart rtorrent? Cheers, -- Norbert Papke. npa...@acm.org http://saveournet.ca Protecting your Internet's level playing field ___ 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-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: ZFS performance degradation over time
On Tue, Jan 19, 2010 at 11:40:50AM -0500, Garrett Moore wrote: I've been watching my memory usage and I have no idea what is consuming memory as 'Active'. Last night I had around 6500MB 'Active' again, 1500MB Wired, no inact, ~30MB buf, no free, and ~100MB swap used. My performance copying ZFS-ZFS was again slow (1MB/s). I tried killing rTorrent and no significant amount of memory was reclaimed - maybe 100MB. `ps aux` showed no processes using any significant amount of memory, and I was definitely nowhere near 6500MB usage. I tried running the perl oneliner again to hog a bunch of memory, and almost all of the Active memory was IMMEDIATELY marked as Free, and my performance was excellent again. I'm not sure what in userland could be causing the issue. The only things I've installed are rTorrent, lighttpd, samba, smartmontools, vim, bash, Python, Perl, and SABNZBd. There is nothing that *should* be consuming any serious amount of memory. I've two recommendations: 1) Have you considered upgrading to RELENG_8 (e.g. 8.0-STABLE) instead of sticking with 8.0-RELEASE? There's been a recent MFC to RELENG_8 which pertain to ARC drainage. I'm referring to the commit labelled revision 1.22.2.2 (RELENG_8): http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c 2) Have you tried using vfs.zfs.arc_max in loader.conf to limit the ARC size? I'd recommend picking something like 1GB as a cap (your machine has 8GB total at present, if I remember right). I believe long ago someone said this isn't an explicit hard limit on the maximum size of the ARC, but I believe this was during the RELENG_7 days and the ARC stuff on FreeBSD has changed since then. I wish the tunables were better documented, or at least explained in detail (hello Wiki!). Finally: Does anyone have reservations about me crossposting this thread to freebsd...@freebsd.org, to get some additional eyes? Reports like this often scare/worry those of us who run servers. :-) On Mon, Jan 18, 2010 at 11:29 AM, Norbert Papke npa...@acm.org wrote: On January 17, 2010, Garrett Moore wrote: I upgraded my system to 8GB of ram to see if that would help. It hasn't made much of a difference. After having rTorrent running for a while, my performance again tanked. Around 6.5GB of memory was showing as 'Active' according to top. 6.5GB of active memory seems to imply that a user process is growing or a large number of user processes are being created. I would expect ZFS's cache to increase the size of wired memory. Sorry, I have not followed this thread closely. Are you sure that the degradation is ZFS related? Could it be caused by, for instance, a userland memory leak? What happens to active memory when you restart rtorrent? -- | 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
8.0-RELEASE / gpart / GPT / marking a partition as active
It seems that quite a few BIOSes have serious issues booting off disks using GPT partitioning when no partition present is marked as active. See http://www.freebsd.org/cgi/query-pr.cgi?pr=115406cat=bin for a prime example. In 8.0-RELEASE, using gpart, setting a slice as active in MBR partitioning mode is trivial, ie: gpart set -a active -i 1 DISKNAME However, trying to do the same thing with GPT partitioning yields no results: gpart set -a active -i 1 DISKNAME gpart: attrib 'active': Device not configured As a result of this issue, I can configure and make a succesfull install using GPT in 8.0, but I cannot boot off it using my Intel D945GCLF2 board. I have found this discussion from about a month ago: http://www.mail-archive.com/freebsd-stable@freebsd.org/msg106918.html where Robert mentions that gpart set -a active -i 1 is no longer needed in 8-STABLE, because the pmbr will be marked as active during the installation of the bootcode. Is there anything I can do to archieve the same result in 8.0-RELEASE or is installing from a snapshop of 8-STABLE my only option? Thanks. - Sincerely, Dan Naumov ___ 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: About nice(1), renice(8) and ULE scheduler
Jordi Espasa Clofent jespa...@minibofh.org wrote: I've realized that the nice(1) code hasn't been modified for a long time (last code change seems from 4 years ago according the sources). ¿Is the nice(1) behaviour the expected? I mean, ¿Has been the ULE scheduler adapted to nice(1) command or not? nice(1) is a very old command based on old concepts and created in times when SMP didn't exist. So ¿it works correctly when a modern an re-designed scheduler as ULE is used? In fact nice is a very simple program. It only changes the priority value of a process in a POSIX-compliant way. There is no need to change or adapt it; it still works fine in the SMP world and with new schedulers. It's up to the scheduler to interpret and handle the priority values of processes. In other words: The nice(1) tool only attaches a number to a process, nothing more. Only the scheduler knows what that number means. So there's no need to change nice(1). By the way, the source code of nice(1) is almost trivial. Basically it just calls the setpriority(2) and execve(2) syscalls. 99% of the source file consists of the BSD license test, arguments parsing and C syntax overhead. 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 I have stopped reading Stephen King novels. Now I just read C code instead. -- Richard A. O'Keefe ___ 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: scp from livecd (fixit option) cd can not find /usr/bin/ssh
On Tue, 19 Jan 2010 10:57, Johan@ wrote: Hello i use the livecd to save an old pc. But when i try to use scp to copy some data to our network i get the message that /usr/bin/ssh could not be found. This is on 8.0 i386 livecd This is the same for snapshots of stable/7 livecd too. -- jhell ___ 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: scp from livecd (fixit option) cd can not find /usr/bin/ssh
On Tue, Jan 19, 2010 at 12:36:12PM -0500, jhell wrote: On Tue, 19 Jan 2010 10:57, Johan@ wrote: Hello i use the livecd to save an old pc. But when i try to use scp to copy some data to our network i get the message that /usr/bin/ssh could not be found. This is the same for snapshots of stable/7 livecd too. Fixit# which ssh /mnt2/usr/bin/ssh Fixit# scp -S /mnt2/usr/bin/ssh ... Jeff ___ 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: 8.0-RELEASE / gpart / GPT / marking a partition as active
on 19/01/2010 19:11 Dan Naumov said the following: I have found this discussion from about a month ago: http://www.mail-archive.com/freebsd-stable@freebsd.org/msg106918.html where Robert mentions that gpart set -a active -i 1 is no longer needed in 8-STABLE, because the pmbr will be marked as active during It was never ever needed, because it never worked. the installation of the bootcode. Is there anything I can do to archieve the same result in 8.0-RELEASE or is installing from a snapshop of 8-STABLE my only option? People did it using fdisk -a, google should have turned that up. -- Andriy Gapon ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
Hi, it seems that I'm not experiencing this at all... Jeremy Chadwick wrote: WD10EACS - WD Caviar Green, 1TB, 16MB, variable rpm have one (external HDD, always powered on) Model = WDC WD10EAVS-00D7B1 Firmware Version = 01.01A01 Power_On_Hours 5757 Power_Cycle_Count 75 Load_Cycle_Count127 WD10EADS - WD Caviar Green, 1TB, 32MB, variable rpm have four, got them in Sep/09. one is RMAed right now because it had defective sectors in Dec/09 (surprisingly the first WDC I encountered problems so far). All four are used as a RAID-5 configured with my 3ware 9500-4LP. I don't know if the controller requests data from the drives in a very short period of time. I'm MRTGing some SMART attributes which should at least query the drive every minute. The ata idle feature is not supported at all trough the 3ware controllers. For the remaining 3 drives: Model = WDC WD10EADS-00L5B1 Firmware Version = 01.01A01 Power_On_Hours 2538 Power_Cycle_Count 140 Load_Cycle_Count141 Power_On_Hours 2526 Power_Cycle_Count 133 Load_Cycle_Count134 Power_On_Hours 2490 Power_Cycle_Count 126 Load_Cycle_Count128 So - I'm not experiencing this problem at all but I feel that the drives are reacting more slowly than my previous WD2500KS-00MJB0 are. changing a directory and then typing ls takes some time until the result is displayed... I can live with that but it doesn't feel good And - it is not really something I would call fast - but it saves power :) The 3ware RAID-5 over 4 disks (using the EAVS from above as replacement right now until WDC sends a new drive back): Version 1.96 --Sequential Output-- --Sequential Input- --Random- Concurrency 1 -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks-- MachineSize K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP nudel.salatschue 6G89 99 76321 38 29285 21 249 99 111225 33 201.7 13 Latency 106ms 247ms1155ms 63744us 145ms 382ms Version 1.96 --Sequential Create-- Random Create nudel.salatschuesse -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete-- files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP 16 571 6 + +++ 1095 8 617 6 + +++ 611 5 Latency 523ms 61us 125ms 103ms 87us 159ms -- Oliver Lehmann http://www.pofo.de/ http://wishlist.ans-netz.de/ ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
On Tue, 19 Jan 2010, Gerrit Kühn wrote: On Tue, 19 Jan 2010 03:24:49 -0800 Jeremy Chadwick free...@jdc.parodius.com wrote about Re: immense delayed write to file system (ZFS and UFS2), performance issues: JC JC If you want a consumer-edition drive that's better tuned for JC JC server work, you should really be looking at the WD Caviar Black JC JC series or their RE/RE2 series. JC That's exactly what I did. I have WD-RE2 drives here that show JC exactly this problem (RE2/GP)! The model number is WD1000FYPS-01ZKB0. JC I should have been more specific. WD makes RE-series drives which JC don't have GP applied to them; those are what I was referring to. Well, when I bought these drives I was not aware of this issue. Buying a drive intended for 24/7 use in RAID configurations is basically the right idea, I think. From what was written about the GP feature back then I could not anticipate such problems. I would have liked to buy the 2TB drives without GP lately, but they have lead times into April here. So I went for the GP model, which now shows the same problem as the 1TB drive... :-( JC WD1000FYPS - WD RE2-GP, 1TB, 16MB, variable rpm JC WD2002FYPS - WD RE4-GP, 2TB, 64MB, variable rpm JC So which drive models above are experiencing a continual increase in JC SMART attribute 193 (Load Cycle Count)? My guess is that some of the JC WD Caviar Green models, and possibly all of the RE2-GP and RE4-GP JC models are experiencing this problem. I can confirm that the two models above show this problem. Furthermore I can confirm that at least in my setup here this drive type works fine: WD5001ABYS I have some of the RE3 drives sitting around here and will probably try them later. I specifically bought RE3 drives recently because of the whole fiasco regarding raid compatibility. I paid about $20 each more for these in the interest of things just working, since I saw some debate about whether or not the TLER setting can be flipped on the non-RAID drives. FWIW, 4 1TB RE3's in a zfs raidz (an excerpt from bonnie++, block read/write 8G on 4G RAM): Write Read K/sec K/sec 123207 142749 Both zfs and these drives kind of surprised me, that seems pretty good for software raid with parity... Seagate is currently on my blacklist after we had a large number of them fail a year or two in the past year or two. I've had good luck so far with WD RE2 and RE3 drives. I'll probably give seagate another shot in a year or two. C Can anyone here report anything about the fixed firmware from http://support.wdc.com/product/download.asp?groupid=609sid=113lang=en? Does this remedy the problem for the 1TB RE2 drive? JC I say some with regards to WD Caviar Green since I have some which do JC not appear to exhibit the heads/actuator arm moved into the JC landing/park zone. I'm at work right now, but when I get home I can JC verify what models I've used which didn't experience this problem, as JC well as what the manufacturing date and F/W revisions are. I should JC note I don't have said Green drives in use (I use WD1001FALS drives JC now). Thanks for sharing this information. cu Gerrit ___ 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-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: 8.0 stable if_bwi kmod not exist?
On Tuesday 19 January 2010 12:44:16 am wsk wrote: folks, There is not exist if_bwi.ko module in /boot/kernel under 8.0 Stable why? ___ freebsd-...@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org Perhaps in /etc/make.conf, you define modules you want to build and this is a new one to be added? Look at variables MODULES_OVERRIDE and WITHOUT_MODULES. ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
Morgan Wesström wrote: Garrett Moore wrote: The drives being discussed in my related thread (regarding poor performance) are all WD Green drives. I have used wdidle3 to set all of my drive timeouts to 5 minutes. I'll see what sort of difference this makes for performance. Even if it makes no difference to performance, thank you for pointing it out -- my drives have less than 2,000 hours on them and were all over 90,000 load cycles due to this moronic factory setting. Since changing the timeout, they haven't parked (which is what I would expect). You're welcome. I just feel as bad for you as for everyone else who has bought these obviously Windoze optimized harddrives. Unfortunately neither wdidle3 nor an updated firmware is available or functioning on the latest models in the Green series. At least that's what I've read from other people having this issue. WD only claims they don't support Linux and they probably have never heard of FreeBSD. If anyone successfully has fixed their WD15EADS drives this way I'd be interested in hearing from you. One of my drives has 216,000 load cycles accumulated over 8 months. That's one every 2nd minute... and I was hit by the Seagate 7200.11 fiasco too. Running on Samsungs now :-) Keeping this python script running prevents Load_Cycle_Count from incrementing on my WD15EADS drives by forcing a write every 5 seconds (2 drive zfs mirror pool, average of 2 load cycles per minute when the script is not running): import time,os mypool = /zfspool # ^^ Change to your pool! fname = os.path.join(mypool, wd_green_anti_idle.pyfile) c = 0 f = open(fname, w) while True: if c == 100: f.close() f = open(fname, w) c = 0 c += 1 time.sleep(5) f.write(a) f.flush() os.fsync(f.fileno()) -- Erik ___ 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: 8.0-RELEASE / gpart / GPT / marking a partition as active
On 1/19/2010 12:11 PM, Dan Naumov wrote: It seems that quite a few BIOSes have serious issues booting off disks using GPT partitioning when no partition present is marked as active. See http://www.freebsd.org/cgi/query-pr.cgi?pr=115406cat=bin for a prime example. In 8.0-RELEASE, using gpart, setting a slice as active in MBR partitioning mode is trivial, ie: gpart set -a active -i 1 DISKNAME However, trying to do the same thing with GPT partitioning yields no results: gpart set -a active -i 1 DISKNAME gpart: attrib 'active': Device not configured As a result of this issue, I can configure and make a succesfull install using GPT in 8.0, but I cannot boot off it using my Intel D945GCLF2 board. I have found this discussion from about a month ago: http://www.mail-archive.com/freebsd-stable@freebsd.org/msg106918.html where Robert mentions that gpart set -a active -i 1 is no longer needed in 8-STABLE, because the pmbr will be marked as active during the installation of the bootcode. Is there anything I can do to archieve the same result in 8.0-RELEASE or is installing from a snapshop of 8-STABLE my only option? After using gpart to create the GPT (and thus the PMBR and its bootcode), why not simply use fdisk -a -1 DISKNAME to set the PMBR partition active? ___ 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: 8.0-RELEASE / gpart / GPT / marking a partition as active
On 1/19/2010 12:11 PM, Dan Naumov wrote: It seems that quite a few BIOSes have serious issues booting off disks using GPT partitioning when no partition present is marked as active. See http://www.freebsd.org/cgi/query-pr.cgi?pr=115406cat=bin for a prime example. In 8.0-RELEASE, using gpart, setting a slice as active in MBR partitioning mode is trivial, ie: gpart set -a active -i 1 DISKNAME However, trying to do the same thing with GPT partitioning yields no results: gpart set -a active -i 1 DISKNAME gpart: attrib 'active': Device not configured As a result of this issue, I can configure and make a succesfull install using GPT in 8.0, but I cannot boot off it using my Intel D945GCLF2 board. I have found this discussion from about a month ago: http://www.mail-archive.com/freebsd-stable@freebsd.org/msg106918.html where Robert mentions that gpart set -a active -i 1 is no longer needed in 8-STABLE, because the pmbr will be marked as active during the installation of the bootcode. Is there anything I can do to archieve the same result in 8.0-RELEASE or is installing from a snapshop of 8-STABLE my only option? After using gpart to create the GPT (and thus the PMBR and its bootcode), why not simply use fdisk -a -1 DISKNAME to set the PMBR partition active? According to the fdisk output, the partition flag did change from 0 to 80. Can the fdisk: Class not found error showing up at the very end of the procedure of doing fdisk -a -1 DISKNAME be safely ignored? - Dan Naumov ___ 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: 8.0-RELEASE / gpart / GPT / marking a partition as active
on 19/01/2010 23:09 Dan Naumov said the following: After using gpart to create the GPT (and thus the PMBR and its bootcode), why not simply use fdisk -a -1 DISKNAME to set the PMBR partition active? According to the fdisk output, the partition flag did change from 0 to 80. Can the fdisk: Class not found error showing up at the very end of the procedure of doing fdisk -a -1 DISKNAME be safely ignored? Yes, I think so. -- Andriy Gapon ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
2010/1/19 Daniel O'Connor docon...@gsoft.com.au On Tue, 19 Jan 2010, Morgan Wesström wrote: The disks involved don't happen to be Western Digital Green Power disks, do they? The Intelli-Park function in these disks are wrecking havoc with I/O in Linux-land at least, causing massive stalls and iowait through the roof during the 25-30 seconds it takes for the heads to unload after parking. I have two of these disks sitting on my desk now collecting dust... There's this.. http://www.silentpcreview.com/Terabyte_Drive_Fix and you can get the tool at.. http://home.arcor.de/ghostadmin/wdidle3_1_00.zip I am planning to try this out tonight.. I just put my WD5000AACS (it has the same problem) in my Windows PC and did a SMART drive quick self-test with the WD utility (Data Lifeguard Diagnostic for Windows). I also tried this Idle Mode Updade Utility but it did not attach to my drive. So i put it back to my FreeBSD box (8.0-Stable) and recognized that I am not able to decrypt it anymore with geli. It keeps telling: # geli attach -k /etc/keys/keyfile /dev/ada0 geli: Cannot read metadata from /dev/ada0: Invalid argument. A geli backup command failed with geli: MD5 hash mismatch: not a geli provider? I think Windows has messed up something on my disk, but a fdisk dump looks still the same as before: # fdisk /dev/ada0 *** Working on device /dev/ada0 *** parameters extracted from in-core disklabel are: cylinders=969021 heads=16 sectors/track=63 (1008 blks/cyl) Figures below won't work with BIOS for partitions not in cyl 1 parameters to be used for BIOS calculations are: cylinders=969021 heads=16 sectors/track=63 (1008 blks/cyl) fdisk: invalid fdisk partition table found Media sector size is 512 Warning: BIOS sector numbering starts with sector 1 Information from DOS bootblock is: The data for partition 1 is: sysid 165 (0xa5),(FreeBSD/NetBSD/386BSD) start 63, size 976773105 (476939 Meg), flag 80 (active) beg: cyl 0/ head 1/ sector 1; end: cyl 316/ head 15/ sector 63 The data for partition 2 is: UNUSED The data for partition 3 is: UNUSED The data for partition 4 is: UNUSED Are there any things I could try or is all my data gone? Thanks in advance ___ 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: Pack of CAM improvements
On Jan 19, 2010, at 9:12 AM, Alexander Motin wrote: Hi. I've made a patch, that should solve set of problems of CAM ATA and CAM generally. I would like to ask for testing and feedback. What patch does: - It unifies bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices). I've fought many times against delay values like this. They never work well enough. Drivers that have delayed scans should set up their own intrhook to delay the boot until their scan is done. To help this out, CAM should move to its own hook that is guaranteed to run after the normal intrhooks. However, this isn't required. Here's my alternate proposal: - move xpt_config() execution to a new config hook that runs after the normal intrhooks. - For self identifying buses (i.e. anything where device presence is known to the controller), have the SIM notify CAM of each target device, instead of assuming that CAM will scan for it. - Teach USB and whatnot to use a confighook to drive their discovery, instead of estimated timeouts. I'm doing exactly this for the new MPT2SAS driver. Scott ___ 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: An old gripe: Reading via mmap stinks
01/14/10 21:56, Matthew Dillon написав(ла): This would explain why the performance is not as bad as linux but is not as good as a properly pipelined case. For what it may be worth, here are the stats for Solaris as well: * Solaris 8, native, 32-bit binary (using -lcrypto instead of -lmd): mmap: 103.54u 27.18s 2:56.46 74.0% read: 99.12u 40.37s 2:53.06 80.6% * Solaris 10, native, 32-bit binary (using -lcrypto instead of -lmd): mmap: 159.36u 83.23s 5:28.25 73.9% read: 173.50u 104.16s 4:48.30 96.3% * Solaris 10, using the 32-bit binary built on Solaris-8: mmap: 217.74u 101.20s 5:58.89 88.8% All of the read results on Solaris (and earlier on Linux) were obtained from using ``openssl md5 largefile''. Seems like BSD is not the only OS, where the mmap's theoretical promise lags behind the actual offering -- read wins on Solaris overall too, despite being quite a bit more expensive in sys-time. Would still be nice to be the first to deliver... -mi ___ 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: thunderbird3: dies with socket(): Protocol not supported Illegal instruction (core dumped)
On 01/12/10 15:31, Guido Falsi wrote: On Tue, Jan 12, 2010 at 02:30:56PM +0100, Gary Jennejohn wrote: On Mon, 11 Jan 2010 12:54:48 +0100 ohart...@zedat.fu-berlin.de said: ohartman Since friday after the last FreeBSD 8.0-STABLE/amd64 update, thunderbird3 ohartman crashes immmediately or after a view seconds with ohartman socket(): Protocol not supported ohartman Illegal instruction (core dumped) I'm not sure but I suspect you are using custom kernel built without INET6 option. If so, thunderbird3 is depending upon IPv6. Can it be made to not require IPv6? (especially when there is no actual IPv6 connectivity). Seems to be hardcoded all over the place. Looks like it would require major modifications. This is strange. I'm using an 8.0-STABL:E/amd64 compiled yesterday here. Custom kernel without INET6 and thunderbird3 works quite fine. Sorry for the long delay. Well, after recompiling several ports the situation seems to get relxed, but thunderbird3 is crashing on SMP boxes more frequent than on non-SMP FreeBSD 8/amd64 boxes. So far. I tried to rebuild via portmaster every thunderbird-dependencies (ports on which tunderbird3 depends on, if my English isn't clear). No success. I realise those crashes when I try to abort message sending, then thunderbird crahes immediately. Sometimes it vanishes silently, leaving a dead-mark (core dump) behind. No idea what's going wrong. On my private slwo UP box (FBSD 8.0/amd64 STABLE) thunderbird is stable, but really slow ... Regards, Oliver ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
On 01/18/10 21:34, � wrote: O. Hartmann wrote: I realise a strange behaviour of several FreeBSD 8.0-STABLE/amd64 boxes. All boxes have the most recent STABLE. One box is a UP system, two others SMP boxes, one with a Q6600 4-core, another XEON with 2x 4-cores (Dell Poweredge III). Symptome: All boxes have ZFS and UFS2 filesystems. Since two weeks or so, sometimes the I/O performance drops massively when doing 'svn update', 'make world' or even 'make kernel'. It doesn't matter what memory and how many cpu the box has, it get stuck for several seconds and freezing. On the UP box, this is sometimes for 10 - 20 seconds. A very interesting phenomenon is the massively delayed file writing on ZFS filesystems I realise. Editing a file in 'vi' running on one XTerm and having in another Xterminal my shell for compiling this file, it takes sometimes up to 20 seconds to get the file updated after it has been written. It's like having an old, slow NFS connection with long cache delays. These massively delayed file transactions are not necessarely under heavy load, sometimes they occur in a relaxed situation. They seem to occur much more often on the UP box than on the SMP boxes, but this strange phenomenon also occur on the Dell Poweredge II, which has 16GB RAM and summa summarum 16 cores. This phenomenon does occur on ZFS- and UFS2 filesystems as well. It is hardly reproducable. Is there any known issue? Ragrds, Oliver The disks involved don't happen to be Western Digital Green Power disks, do they? The Intelli-Park function in these disks are wrecking havoc with I/O in Linux-land at least, causing massive stalls and iowait through the roof during the 25-30 seconds it takes for the heads to unload after parking. I have two of these disks sitting on my desk now collecting dust... /Morgan ___ 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 The disks in question are indeed WD on one box, but they are all Caviar Black and they performed well months ago with the very same hardware and an earlier FreeBSD 8 version. The other boxes in questions do have a set of mixed type, Seagate, WD, Samsung (mostly Samsung F1 types). We do not use 'Green' drives, due to every box acts as a server and we found green-disks, even from WD, too slow. Oliver ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
On 01/19/10 10:09, krad wrote: 2010/1/18 Morgan Wesstr�mfreebsd-questi...@pp.dyndns.biz O. Hartmann wrote: I realise a strange behaviour of several FreeBSD 8.0-STABLE/amd64 boxes. All boxes have the most recent STABLE. One box is a UP system, two others SMP boxes, one with a Q6600 4-core, another XEON with 2x 4-cores (Dell Poweredge III). Symptome: All boxes have ZFS and UFS2 filesystems. Since two weeks or so, sometimes the I/O performance drops massively when doing 'svn update', 'make world' or even 'make kernel'. It doesn't matter what memory and how many cpu the box has, it get stuck for several seconds and freezing. On the UP box, this is sometimes for 10 - 20 seconds. A very interesting phenomenon is the massively delayed file writing on ZFS filesystems I realise. Editing a file in 'vi' running on one XTerm and having in another Xterminal my shell for compiling this file, it takes sometimes up to 20 seconds to get the file updated after it has been written. It's like having an old, slow NFS connection with long cache delays. These massively delayed file transactions are not necessarely under heavy load, sometimes they occur in a relaxed situation. They seem to occur much more often on the UP box than on the SMP boxes, but this strange phenomenon also occur on the Dell Poweredge II, which has 16GB RAM and summa summarum 16 cores. This phenomenon does occur on ZFS- and UFS2 filesystems as well. It is hardly reproducable. Is there any known issue? Ragrds, Oliver The disks involved don't happen to be Western Digital Green Power disks, do they? The Intelli-Park function in these disks are wrecking havoc with I/O in Linux-land at least, causing massive stalls and iowait through the roof during the 25-30 seconds it takes for the heads to unload after parking. I have two of these disks sitting on my desk now collecting dust... /Morgan ___ freebsd-questi...@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to freebsd-questions-unsubscr...@freebsd.org ZFS is copy on write, therefore to optimize the write performance it delays writes for a long as possible, upto a set maximum time. It will then flush to the disks. How long this time is depends on how much free ram you have available. Assuming processes are eating up all your ram I would imagine you are hitting the max limit. I'm not sure exactly what its set to on bsd but I know the default on opensolaris is 30s. I think this explains your delayed writes. Not sure what will cause the lock ups though. ___ 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 This could end in a bad situation, where one process writes a files, say with some arbitrary stuff and another successing process is intended to read this file. even if the processes are run serial, those 'delays' could break the chain! The delay situation in a development environment is harsh, but in other circumstances it could develop very bad. I see this strange behaviour now for several weeks, something essential has changed in the code, I guess. On UP boxes the situation is worse sometimes, on SMp boxes with lots of RAM ( 8 and 16 GB and 4 or 8 CPU cores) it is still bad. I have a server that acts as a 'rsync' backup system gathering data from satellite servers from time to time. Since this problem of slowness occured, this 4-core 8 gig RAM box crawls for minutes. Even when X11 is disabled working on console is 'bumpy': terminal out slows down, mouse pointer jumps etc.As I wrote, the same on a 8 core/16 gig box, but not that harsh. ___ 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: 8.0 stable if_bwi kmod not exist?
于 2010/01/20 02:57, Steven Friedrich 写道: On Tuesday 19 January 2010 12:44:16 am wsk wrote: folks, There is not exist if_bwi.ko module in /boot/kernel under 8.0 Stable why? ___ freebsd-...@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org Perhaps in /etc/make.conf, you define modules you want to build and this is a new one to be added? Look at variables MODULES_OVERRIDE and WITHOUT_MODULES. but not any define in my make.conf. it is bizarre # added by use.perl 2010-01-18 15:02:07 PERL_VERSION=5.8.9 ___ 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: 8.0 stable if_bwi kmod not exist?
On Tue, Jan 19, 2010 at 01:44:16PM +0800, wsk wrote: folks, There is not exist if_bwi.ko module in /boot/kernel under 8.0 Stable why? Looks like it wasn't hooked up to the build for some reason, no bwi here: http://svn.freebsd.org/viewvc/base/stable/8/sys/modules/Makefile?revision=202410view=markup HTH, Yuri ___ 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: immense delayed write to file system (ZFS and UFS2), performance issues
--On Wednesday, January 20, 2010 1:16 AM +0100 O. Hartmann ohart...@mail.zedat.fu-berlin.de wrote: This could end in a bad situation, where one process writes a files, say with some arbitrary stuff and another successing process is intended to read this file. even if the processes are run serial, those 'delays' could break the chain! The delay situation in a development environment is harsh, but in other circumstances it could develop very bad. The read occurs from the write cache. Thus the reader would see the writers data (given the usual POSIX semantics). ZFS delayed writes are handled by the cache/ZIL layers, reads and writes go through these layers. The ZIL is normally written very quickly. ZFS actually (through the ZIL) has better journalling and recovery semantics than most journalling filesystems. I see this strange behaviour now for several weeks, something essential has changed in the code, I guess. On UP boxes the situation is worse sometimes, on SMp boxes with lots of RAM ( 8 and 16 GB and 4 or 8 CPU cores) it is still bad. I have a server that acts as a 'rsync' backup system gathering data from satellite servers from time to time. Since this problem of slowness occured, this 4-core 8 gig RAM box crawls for minutes. Even when X11 is disabled working on console is 'bumpy': terminal out slows down, mouse pointer jumps etc.As I wrote, the same on a 8 core/16 gig box, but not that harsh. ___ 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-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: 8.0-RELEASE / gpart / GPT / marking a partition as active
On 1/19/10 4:09 PM, Dan Naumov wrote: On 1/19/2010 12:11 PM, Dan Naumov wrote: It seems that quite a few BIOSes have serious issues booting off disks using GPT partitioning when no partition present is marked as active. See http://www.freebsd.org/cgi/query-pr.cgi?pr=115406cat=bin for a prime example. In 8.0-RELEASE, using gpart, setting a slice as active in MBR partitioning mode is trivial, ie: gpart set -a active -i 1 DISKNAME However, trying to do the same thing with GPT partitioning yields no results: gpart set -a active -i 1 DISKNAME gpart: attrib 'active': Device not configured As a result of this issue, I can configure and make a succesfull install using GPT in 8.0, but I cannot boot off it using my Intel D945GCLF2 board. I have found this discussion from about a month ago: http://www.mail-archive.com/freebsd-stable@freebsd.org/msg106918.html where Robert mentions that gpart set -a active -i 1 is no longer needed in 8-STABLE, because the pmbr will be marked as active during the installation of the bootcode. Is there anything I can do to archieve the same result in 8.0-RELEASE or is installing from a snapshop of 8-STABLE my only option? After using gpart to create the GPT (and thus the PMBR and its bootcode), why not simply use fdisk -a -1 DISKNAME to set the PMBR partition active? According to the fdisk output, the partition flag did change from 0 to 80. Can the fdisk: Class not found error showing up at the very end of the procedure of doing fdisk -a -1 DISKNAME be safely ignored? Yes, just ignore it. ___ 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: Multiple serial consoles via null modem cable
Hello, Using `cu' only works with COM1 for me. Currently I have two serial ports on the system, and only the first is able to make the connection - the serial consoles are enabled in /etc/tty, but as I said only COM1 is able to make the connection. Regards, Marin On Sun, Jan 17, 2010 at 3:32 PM, Ronald Klop ronald-freeb...@klop.yi.orgwrote: On Fri, 15 Jan 2010 07:34:17 +0100, Marin Atanasov dna...@gmail.com wrote: Thank you a lot for your feedback! Now to the real question again, because I'm a little confused now - can I still get a usb-to-serial port converter having let's say 8 serial ports and then connect each machine to the usb-to-serial hub and manage them remotely from a single location (the host having the usb-to-serial hub)? That way I just specify a serial port number and I get to a specific machine? The model provided by Boris looks nice, and that was my initial idea, but I'm not sure if I could get it working under FreeBSD. Is conserver or conserver-com able to handle this? I know that cu uses COM1 only, but will conserver able to handle serial consoles on different ports, since the usb-to-serial port would appear as multiple serial ports. You can provide cu with the port to connect to on the command line. cu -l cuaU0 -s 115200 cu -l cuaU1 -s 115200 etc. You can not connect several servers on 1 serial port, but you can connect several servers on several serial ports. With serial-over-usb it scales to many serial ports. Ronald. Thank you and regards, Marin On Tue, Jan 12, 2010 at 7:04 PM, Boris Samorodov b...@ipt.ru wrote: On Tue, 12 Jan 2010 17:14:44 +0200 Marin Atanasov wrote: I'm thinking about the following situation - 1 system acting like a host with a serial port hub, each port of the hub is connected to a different machine on sio0, using null modem cables. Along with milti-io serial cards we use multi-usb serial converters, such as SUNIX UTS7009P (7 USB to serial adapter): http://www.sunix.com.tw/it/en/LinkCraft/UTS4009P_UTS7009P.htm -- WBR, Boris Samorodov (bsam) Research Engineer, http://www.ipt.ru Telephone Internet SP FreeBSD Committer, http://www.FreeBSD.org The Power To Serve -- Marin Atanasov Nikolov dnaeon AT gmail DOT com daemon AT unix-heaven DOT org ___ 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
top Segmentation faulting on 8.0p2 amd64
Dear all, I have no idea why top crashes with segmentation fault on my amd64 machine running FreeBSD 8.0-RELEASE-p2. If someone wants to have a loot at the core dump: http://www.schmalzbauer.de/downloads/top.core But I think I should recompile it with DEBUG=-g first, right? World and kernel are in sync, I ran an extra build to make that sure. My kernconf follows. Any help appreciated. Thanks, FreeBSD qaweb.hn.sand.spsnetz.de 8.0-RELEASE-p2 FreeBSD 8.0-RELEASE-p2 #0: Tue Jan 19 22:00:17 CET 2010 ad...@qaweb.hn.sand.spsnetz.de:/usr/obj/usr/src/sys/FJS-RX100S3 amd64 include GENERIC makeoptions none nooptions COMPAT_FREEBSD4 # Compatible with FreeBSD4 nooptions COMPAT_FREEBSD5 # Compatible with FreeBSD5 # Floppy drives nodevicefdc # ATA and ATAPI devices #device ata #device atadisk # ATA disk drives nodeviceataraid # ATA RAID drives #device atapicd # ATAPI CDROM drives nodeviceatapifd # ATAPI floppy drives nodeviceatapist # ATAPI tape drives #optionsATA_STATIC_ID # Static device numbering # SCSI Controllers nodeviceahb # EISA AHA1742 family nodeviceahc # AHA2940 and onboard AIC7xxx devices nooptions AHC_REG_PRETTY_PRINT# Print register bitfields in debug # output. Adds ~128k to driver. nodeviceahd # AHA39320/29320 and onboard AIC79xx devices nooptions AHD_REG_PRETTY_PRINT# Print register bitfields in debug # output. Adds ~215k to driver. nodeviceamd # AMD 53C974 (Tekram DC-390(T)) nodevicehptiop # Highpoint RocketRaid 3xxx series nodeviceisp # Qlogic family ##deviceispfw # Firmware for QLogic HBAs- normally a module nodevicempt # LSI-Logic MPT-Fusion ##devicencr # NCR/Symbios Logic nodevicesym # NCR/Symbios Logic (newer chipsets + those of `ncr') nodevicetrm # Tekram DC395U/UW/F DC315U adapters nodeviceadv # Advansys SCSI adapters nodeviceadw # Advansys wide SCSI adapters nodeviceaha # Adaptec 154x SCSI adapters nodeviceaic # Adaptec 15[012]x SCSI adapters, AIC-6[23]60. nodevicebt # Buslogic/Mylex MultiMaster SCSI adapters nodevicencv # NCR 53C500 nodevicensp # Workbit Ninja SCSI-3 nodevicestg # TMC 18C30/18C50 # SCSI peripherals device scbus # SCSI bus (required for SCSI) device ch # SCSI media changers #device da # Direct Access (disks) #device sa # Sequential Access (tape etc) #device cd # CD #device pass# Passthrough device (direct SCSI access) #device ses # SCSI Environmental Services (and SAF-TE) # RAID controllers interfaced to the SCSI subsystem nodeviceamr # AMI MegaRAID nodevicearcmsr # Areca SATA II RAID ##XXX it is not 64-bit clean, -scottl ##deviceasr # DPT SmartRAID V, VI and Adaptec SCSI RAID nodeviceciss# Compaq Smart RAID 5* nodevicedpt # DPT Smartcache III, IV - See NOTES for options nodevicehptmv # Highpoint RocketRAID 182x nodevicehptrr # Highpoint RocketRAID 17xx, 22xx, 23xx, 25xx nodeviceiir # Intel Integrated RAID nodeviceips # IBM (Adaptec) ServeRAID nodevicemly # Mylex AcceleRAID/eXtremeRAID nodevicetwa # 3ware 9000 series PATA/SATA RAID # RAID controllers nodeviceaac # Adaptec FSA RAID nodeviceaacp# SCSI passthrough for aac (requires CAM) nodeviceida # Compaq Smart RAID nodevicemfi # LSI MegaRAID SAS nodevicemlx # Mylex DAC960 family #XXX pointer/int warnings ##devicepst # Promise Supertrak SX6000 nodevicetwe # 3ware ATA RAID # atkbdc0 controls both the keyboard and the PS/2 mouse #device atkbdc # AT keyboard controller #device atkbd # AT keyboard #device psm # PS/2 mouse nodevicekbdmux # keyboard multiplexer #device vga # VGA video card
Re: Pack of CAM improvements
Scott Long wrote: On Jan 19, 2010, at 9:12 AM, Alexander Motin wrote: I've made a patch, that should solve set of problems of CAM ATA and CAM generally. I would like to ask for testing and feedback. What patch does: - It unifies bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices). I've fought many times against delay values like this. They never work well enough. Drivers that have delayed scans should set up their own intrhook to delay the boot until their scan is done. To help this out, CAM should move to its own hook that is guaranteed to run after the normal intrhooks. However, this isn't required. I am sure that using delay is not a perfect solution, but it nicely fits new scanning procedure and costs just a few lines of code. Nobody denies to add _more_ events to wait there. This is just a band-aid for cases when nothing else helps. May be I am mixing something, but AFAIR there were some USB devices, which doesn't appear on a first bus scan, but register later. Here's my alternate proposal: - move xpt_config() execution to a new config hook that runs after the normal intrhooks. To make scanning work in background, it is better to call xpt_config() same as now, as early as possible, to start scanning for already registered buses, but call xpt_release_boot() on some later event (or even several different events). - For self identifying buses (i.e. anything where device presence is known to the controller), have the SIM notify CAM of each target device, instead of assuming that CAM will scan for it. Nobody denies your driver to call xpt_rescan on per-known-device basis. In such case CAM will still wait until all of your scan requests will be fulfilled. You may see it is already done by some SIMs and PMP driver. If you need a way to avoid full scan, it also can be done, while it is separate question. - Teach USB and whatnot to use a confighook to drive their discovery, instead of estimated timeouts. I'm doing exactly this for the new MPT2SAS driver. -- Alexander Motin ___ 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: Multiple serial consoles via null modem cable
On Wed, Jan 20, 2010 at 08:46:48AM +0200, Marin Atanasov wrote: Hello, Using `cu' only works with COM1 for me. Currently I have two serial ports on the system, and only the first is able to make the connection - the serial consoles are enabled in /etc/tty, but as I said only COM1 is able to make the connection. I'm a little confused by this statement, so I'll add some clarify: /etc/ttys is for configuring a machine to tie getty (think login prompt) to a device (in this case, a serial port). Meaning: the device on the other end of the serial cable will start seeing login: and so on assuming you attach to the serial port there. For example: box1 COM1/ttyu0 is wired to box2 COM3/ttyu2 using a null modem cable. box1 COM2/ttyu1 is wired to box2 COM4/ttyu3 using a null modem cable. On box1, you'd have something like this in /etc/ttys: ttyu0 /usr/libexec/getty std.9600 vt100 on secure ttyu1 /usr/libexec/getty std.9600 vt100 on secure This means that login prompts for box1 will be spawned/available on both serial ports (ttyu0 and ttyu1). If you get on box2 and do cu -l ttyu2, this will connect you to box2's COM3 port, which is physically connected to box1's COM1 port. Hit enter and you should see a login: prompt for box1. The same applies if you get on box2 and do cu -l ttyu3 (but for box2's COM4 port, which is wired to box1's COM2 port). With the above configuration in mind, you SHOULD NOT: - Mess with /etc/ttys on box2 - Execute cu -l ttyu0 or cu -l ttyu1 on box1 -- this probably won't work (likely will return some message about the device being locked or in use already). You cannot do something like where box1 COM1 is wired to box2 COM1, and depending on what box you're on doing the cu -l ttyu0 from, get a login prompt on the other. It doesn't work like that. :-) Now, about actual *serial console* itself -- that is to say, kernel output during boot, etc... on a serial port. AFAIK, on FreeBSD you can only set serial console to a single serial port, and that defaults to COM1/ttyu0. You can change what port/device, but there can only be one. HTH... On Sun, Jan 17, 2010 at 3:32 PM, Ronald Klop ronald-freeb...@klop.yi.orgwrote: On Fri, 15 Jan 2010 07:34:17 +0100, Marin Atanasov dna...@gmail.com wrote: Thank you a lot for your feedback! Now to the real question again, because I'm a little confused now - can I still get a usb-to-serial port converter having let's say 8 serial ports and then connect each machine to the usb-to-serial hub and manage them remotely from a single location (the host having the usb-to-serial hub)? That way I just specify a serial port number and I get to a specific machine? The model provided by Boris looks nice, and that was my initial idea, but I'm not sure if I could get it working under FreeBSD. Is conserver or conserver-com able to handle this? I know that cu uses COM1 only, but will conserver able to handle serial consoles on different ports, since the usb-to-serial port would appear as multiple serial ports. You can provide cu with the port to connect to on the command line. cu -l cuaU0 -s 115200 cu -l cuaU1 -s 115200 etc. You can not connect several servers on 1 serial port, but you can connect several servers on several serial ports. With serial-over-usb it scales to many serial ports. Ronald. Thank you and regards, Marin On Tue, Jan 12, 2010 at 7:04 PM, Boris Samorodov b...@ipt.ru wrote: On Tue, 12 Jan 2010 17:14:44 +0200 Marin Atanasov wrote: I'm thinking about the following situation - 1 system acting like a host with a serial port hub, each port of the hub is connected to a different machine on sio0, using null modem cables. Along with milti-io serial cards we use multi-usb serial converters, such as SUNIX UTS7009P (7 USB to serial adapter): http://www.sunix.com.tw/it/en/LinkCraft/UTS4009P_UTS7009P.htm -- | 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
