Re: ZFS performance on 7.2-release/amd64 low compared to UFS2 + SoftUpdates
All the ZFS tuning guides for FreeBSD (including one on the FreeBSD ZFS wiki) have recommended values between 64M and 128M to improve stability, so that what I went with. How much of my max kmem is it safe to give to ZFS? If you are on amd64 then don't tune it, it will tune itself. If you are on i386 (or an earlier verions of amd64) then 128M on a 2 gig machine should be OK, assuming you have kmem_size_max set to the full 1500 odd. Those are numbers which come up time and time again - I ran reliably with them for ages, until the latest -STABLE. -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
Upgrade from 7.1-RELEASE to 7.2-RELEASE through freebsd update
I tried to upgrade my 7.1-RELEASE into 7.2-RELEASE. However freebsd-update kept asking me to merge every file in /etc whose $Id$ line changed (that makes about all files). Is there a way, as with mergemaster, to make it not consider the $Id$ line for the manual merge ? Thank you. -- Erwan ___ 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: Upgrade from 7.1-RELEASE to 7.2-RELEASE through freebsd update
On Thu, Jun 18, 2009 at 10:34:54AM +0200, Erwan David typed: I tried to upgrade my 7.1-RELEASE into 7.2-RELEASE. However freebsd-update kept asking me to merge every file in /etc whose $Id$ line changed (that makes about all files). Is there a way, as with mergemaster, to make it not consider the $Id$ line for the manual merge ? You could let freebsd-update ignore /etc (but not /usr/src) and use mergemaster for your configuration files. Ruben ___ 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: panic on 6.4-R in ioapic_get_vector() during device probe
On Wednesday 17 June 2009 8:13:31 am pluknet wrote: Hi. This is on 6.4-RELEASE-p5 Early in boot (probably due to network outage):: Hit [Enter] to boot immediately, or any other key for command prompt. Booting [/boot/kernel/kernel]... /boot/kernel/acpi.ko text*0x44f40 | readin failed elf32*loadimage: read failed GDB: no debug ports present and then.. Timecounter i8254 frequency 1193182 Hz quality 0 CPU: Intel(R) Xeon(R) CPU E5440 @ 2.83GHz (2826.26-MHz 686-class CPU) Origin = GenuineIntel Id = 0x1067a Stepping = 10 Features=0xbfebfbffFPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,C MOV,PAT,PSE36,CLFLUSH,DTS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,PBE Features2=0x40ce3bdSSE3,RSVD2,MON,DS_CPL,VMX,EST,TM2,SSSE3,CX16,xTPR,PDCM,DCA ,b19,b26 AMD Features=0x2000LM AMD Features2=0x1LAHF Cores per package: 4 real memory = 3220992000 (3071 MB) avail memory = 3150835712 (3004 MB) FreeBSD/SMP: Multiprocessor System Detected: 8 CPUs cpu0 (BSP): APIC ID: 0 cpu1 (AP): APIC ID: 1 cpu2 (AP): APIC ID: 2 cpu3 (AP): APIC ID: 3 cpu4 (AP): APIC ID: 4 cpu5 (AP): APIC ID: 5 cpu6 (AP): APIC ID: 6 cpu7 (AP): APIC ID: 7 user VMEM accounting on ioapic0: Assuming intbase of 0 MPTable: Ignoring interrupt entry for missing ioapic0 ioapic0 Version 2.0 irqs 0-23 on motherboard The 'ignoring interrupt entry' message is very odd. Can you get output from 'mptable'? Are you able to boot with ACPI enabled? At this point I would not be surprised if the MP Table was just flat wrong on modern machines as it seems many BIOS vendors do not test it anymore but only test the ACPI tables. -- John Baldwin ___ 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: panic on 6.4-R in ioapic_get_vector() during device probe
2009/6/18 John Baldwin j...@freebsd.org: On Wednesday 17 June 2009 8:13:31 am pluknet wrote: Hi. This is on 6.4-RELEASE-p5 Early in boot (probably due to network outage):: Hit [Enter] to boot immediately, or any other key for command prompt. Booting [/boot/kernel/kernel]... /boot/kernel/acpi.ko text*0x44f40 | readin failed elf32*loadimage: read failed GDB: no debug ports present and then.. Timecounter i8254 frequency 1193182 Hz quality 0 CPU: Intel(R) Xeon(R) CPU E5440 @ 2.83GHz (2826.26-MHz 686-class CPU) Origin = GenuineIntel Id = 0x1067a Stepping = 10 Features=0xbfebfbffFPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,C MOV,PAT,PSE36,CLFLUSH,DTS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,PBE Features2=0x40ce3bdSSE3,RSVD2,MON,DS_CPL,VMX,EST,TM2,SSSE3,CX16,xTPR,PDCM,DCA ,b19,b26 AMD Features=0x2000LM AMD Features2=0x1LAHF Cores per package: 4 real memory = 3220992000 (3071 MB) avail memory = 3150835712 (3004 MB) FreeBSD/SMP: Multiprocessor System Detected: 8 CPUs cpu0 (BSP): APIC ID: 0 cpu1 (AP): APIC ID: 1 cpu2 (AP): APIC ID: 2 cpu3 (AP): APIC ID: 3 cpu4 (AP): APIC ID: 4 cpu5 (AP): APIC ID: 5 cpu6 (AP): APIC ID: 6 cpu7 (AP): APIC ID: 7 user VMEM accounting on ioapic0: Assuming intbase of 0 MPTable: Ignoring interrupt entry for missing ioapic0 ioapic0 Version 2.0 irqs 0-23 on motherboard The 'ignoring interrupt entry' message is very odd. Can you get output from 'mptable'? I'm afraid that panic was only once and due to acpi.ko network load problem. I can boot this box with acpi opted out explicitly if it makes sense, also in order to reproduce those conditions. Are you able to boot with ACPI enabled? Of course. These boxes boot always fine with ACPI enabled. Below is part of related dmesg (now from from 7.2) with ACPI enabled: --- FreeBSD/SMP: Multiprocessor System Detected: 8 CPUs cpu0 (BSP): APIC ID: 0 cpu1 (AP): APIC ID: 1 cpu2 (AP): APIC ID: 2 cpu3 (AP): APIC ID: 3 cpu4 (AP): APIC ID: 4 cpu5 (AP): APIC ID: 5 cpu6 (AP): APIC ID: 6 cpu7 (AP): APIC ID: 7 This module (opensolaris) contains code covered by the Common Development and Distribution License (CDDL) see http://opensolaris.org/os/licensing/opensolaris_license/ ioapic0 Version 2.0 irqs 0-23 on motherboard --- At this point I would not be surprised if the MP Table was just flat wrong on modern machines as it seems many BIOS vendors do not test it anymore but only test the ACPI tables. : # mptable === MPTable --- MP Floating Pointer Structure: location: EBDA physical address: 0x0009ad40 signature:'_MP_' length: 16 bytes version: 1.4 checksum: 0xc9 mode: Virtual Wire --- MP Config Table Header: physical address: 0x0009be10 signature:'PCMP' base table length:716 version: 1.4 checksum: 0xd6 OEM ID: 'IBM ENSW' Product ID: 'x3650 SMP ' OEM table pointer:0x OEM table size: 0 entry count: 72 local APIC address: 0xfee0 extended table length:328 extended table checksum: 217 --- MP Config Base Table Entries: -- Processors: APIC ID Version State Family Model StepFlags 0 0x14BSP, usable 6 7 10 0x0301 1 0x14AP, usable 6 7 10 0x0301 2 0x14AP, usable 6 7 10 0x0301 3 0x14AP, usable 6 7 10 0x0301 4 0x14AP, usable 6 7 10 0x0301 5 0x14AP, usable 6 7 10 0x0301 6 0x14AP, usable 6 7 10 0x0301 7 0x14AP, usable 6 7 10 0x0301 -- Bus:Bus ID Type 0 PCI 1 PCI 2 PCI 3 PCI 4 PCI 5 PCI 6 PCI 7 PCI 8 PCI 9 PCI 10 PCI 11 PCI 12 PCI 13 PCI 14 PCI 15 PCI 16 PCI 17
Re: Issues with gjournal (heaaaaaaaaaaalp!)
On Wed, 10 Jun 2009 23:24:49 -0700 Garrett Cooper yanef...@gmail.com wrote: Hi Dan, I'm doing that right now =\... orangebox# mount /dev/ad6s1a on / (ufs, local) devfs on /dev (devfs, local) /dev/ad6s1d.journal on /usr (ufs, asynchronous, local) /dev/ad6s1e on /usr/home (ufs, local) /dev/ad6s1f on /var (ufs, local) Thanks! -Garrett GJournal actually doesn't work on your box now. To make it work, you MUST use special flag -J to newfs. See beginning of newfs(8). After you do newfs -O2 -J /dev/ad6s1d.journal and mount it, you will see somehow different mount output: = ki...@kibab-nb%mount /dev/ad4s1a on / (ufs, local) devfs on /dev (devfs, local) snip /dev/label/Home on /home (ufs, asynchronous, local, gjournal) = Notice gjournal flag in brackets. You output doesn't contain it... -- Ilya Bakulin webmas...@kibab.com xmpp://kibab...@jabber.ru pgpYe6uUKJlB6.pgp Description: PGP signature
Re: panic on 6.4-R in ioapic_get_vector() during device probe
On Thursday 18 June 2009 10:05:14 am pluknet wrote: 2009/6/18 John Baldwin j...@freebsd.org: On Wednesday 17 June 2009 8:13:31 am pluknet wrote: Hi. This is on 6.4-RELEASE-p5 Early in boot (probably due to network outage):: Hit [Enter] to boot immediately, or any other key for command prompt. Booting [/boot/kernel/kernel]... /boot/kernel/acpi.ko text*0x44f40 | readin failed elf32*loadimage: read failed GDB: no debug ports present and then.. Timecounter i8254 frequency 1193182 Hz quality 0 CPU: Intel(R) Xeon(R) CPU E5440 @ 2.83GHz (2826.26-MHz 686-class CPU) Origin = GenuineIntel Id = 0x1067a Stepping = 10 Features=0xbfebfbffFPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,C MOV,PAT,PSE36,CLFLUSH,DTS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,PBE Features2=0x40ce3bdSSE3,RSVD2,MON,DS_CPL,VMX,EST,TM2,SSSE3,CX16,xTPR,PDCM,DCA ,b19,b26 AMD Features=0x2000LM AMD Features2=0x1LAHF Cores per package: 4 real memory = 3220992000 (3071 MB) avail memory = 3150835712 (3004 MB) FreeBSD/SMP: Multiprocessor System Detected: 8 CPUs cpu0 (BSP): APIC ID: 0 cpu1 (AP): APIC ID: 1 cpu2 (AP): APIC ID: 2 cpu3 (AP): APIC ID: 3 cpu4 (AP): APIC ID: 4 cpu5 (AP): APIC ID: 5 cpu6 (AP): APIC ID: 6 cpu7 (AP): APIC ID: 7 user VMEM accounting on ioapic0: Assuming intbase of 0 MPTable: Ignoring interrupt entry for missing ioapic0 ioapic0 Version 2.0 irqs 0-23 on motherboard The 'ignoring interrupt entry' message is very odd. Can you get output from 'mptable'? I'm afraid that panic was only once and due to acpi.ko network load problem. I can boot this box with acpi opted out explicitly if it makes sense, also in order to reproduce those conditions. Ah, ok. Your MP Table is just plain busted so you will get this panic if you boot with ACPI disabled. You can just compile ACPI into your kernel to ensure you never boot without it. -- John Baldwin ___ 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 on 7.2-release/amd64 low compared to UFS2 + SoftUpdates
On Thu, Jun 18, 2009 at 1:29 AM, Pete French petefre...@ticketswitch.comwrote: All the ZFS tuning guides for FreeBSD (including one on the FreeBSD ZFS wiki) have recommended values between 64M and 128M to improve stability, so that what I went with. How much of my max kmem is it safe to give to ZFS? If you are on amd64 then don't tune it, it will tune itself. If you are on i386 (or an earlier verions of amd64) then 128M on a 2 gig machine should be OK, assuming you have kmem_size_max set to the full 1500 odd. Those are numbers which come up time and time again - I ran reliably with them for ages, until the latest -STABLE. My rule of thumb for 32-bit i386 systems has been to: - assign half of RAM to kmem (up to the max of ~1500 on 7.0/7.1) - assign half of kmem to zfs_arc_max So far, for my workloads (nfs/cifs file servers, cups print servers, rsync servers, kde4 desktop), it's worked well. -- Freddie Cash fjwc...@gmail.com ___ 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
ZFS user library?
Hello, I was wondering if there are plans to document and keep the ZFS user library as a reasonably stable API. I have been writing an automatic replication program, and it's ugly and clumsy to do it calling a user program. I would rather prefer to use an API, that would make it much easier to retrieve error messages, etc. Borja. ___ 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 user library?
I was wondering if there are plans to document and keep the ZFS user library as a reasonably stable API. You really need to ask that on the ZFS lists. Usually Solaris man pages indicate that an API is not stable (assuming) man pages exist. With a few minor exceptions, ZFS in FreeBSD just tracks ZFS in OpenSolaris. Cheers, Kip ___ 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
xorg and intel driver
hail, I know this was here before, http://lists.freebsd.org/pipermail/freebsd-current/2009-March/004775.html, but there was no happy ending there ... is there any news ? I have a STABLE from yesterday and the xorg is too much slow. xorg is from 7.2R cdrom, intel video driver is from today. card is vgap...@pci0:0:2:0: class=0x03 card=0x50448086 chip=0x29c28086 rev=0x02 hdr=0x00 vendor = 'Intel Corporation' device = 'P35/G33 (Bearlake) Integrated Graphics Controller' class = display subclass = VGA if more info is needed, thanks, matheus -- We will call you cygnus, The God of balance you shall be A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? http://en.wikipedia.org/wiki/Posting_style ___ 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: Upgrade from 7.1-RELEASE to 7.2-RELEASE through freebsd update
Le Thu 18/06/2009, Ruben de Groot disait On Thu, Jun 18, 2009 at 10:34:54AM +0200, Erwan David typed: I tried to upgrade my 7.1-RELEASE into 7.2-RELEASE. However freebsd-update kept asking me to merge every file in /etc whose $Id$ line changed (that makes about all files). Is there a way, as with mergemaster, to make it not consider the $Id$ line for the manual merge ? You could let freebsd-update ignore /etc (but not /usr/src) and use mergemaster for your configuration files. Ruben I do not master enough either freebsd-update nor mergemaster to tell the former not to do the merge, but only download what is necessary for the latter to work. If I understand well, I can tell not to touch /etc (removing it from MergeChanges), but I would not be able to find what is then needed for mergemaster. -- Erwan ___ 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: mini-HEADSUP bce owners: please test
We have a Dell R710 with its first ethernet port connected to a Dell 5524 ethernet switch. Just installed 7.2-RELEASE on it and get no connectivity over the ethernet. tcpdump shows no frames at all. Installing your patches eliminated the PHY write timeout errors, but still have no connectivity. Now, upon boot and a couple times thereafter, we see these errors: bce0: ../../../dev/bce/if_bce.c(6968): Watchdog timeout occurred, resetting! bce0: ../../../dev/bce/if_bce.c(1386); Unable to write CTX memory: cid_addr = 0x, offset = 0x! bce0: ../../../dev/bce/if_bce.c(1386); Unable to write CTX memory: cid_addr = 0x, offset = 0x0010! (and so on, for offsets 14, 80, 240, 258, and 25C). ___ 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 user library?
On Jun 18, 2009, at 13:21, Kip Macy wrote: I was wondering if there are plans to document and keep the ZFS user library as a reasonably stable API. You really need to ask that on the ZFS lists. Usually Solaris man pages indicate that an API is not stable (assuming) man pages exist. With a few minor exceptions, ZFS in FreeBSD just tracks ZFS in OpenSolaris. As mentioned above, there is a libzfs but the Sun people are still changing things a lot so they can't guarantee compatibility. One example of these changes is the crypto work being done in OpenSolaris: http://www.opensolaris.org/os/project/zfs-crypto/phase1/libzfs_api/ Is there something specific you're looking to do? The file system layer of ZFS (the ZPL) is in flux, but there may be other components (e.g., DMU) that may be more stable (the Lustre folks are coding against it in user land). See pages 7 and 8 for the three main layers: http://opensolaris.org/os/community/zfs/docs/zfs_last.pdf ___ 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: 6.2 sporadically locks up
Just modify the driver slightly to hijack a different device prefix :) Adrian 2009/6/17 pluknet pluk...@gmail.com: 2009/6/17 Ed Maste ema...@freebsd.org: On Tue, Jun 16, 2009 at 07:03:34PM +0400, pluknet wrote: As for allpcpu, I often see the picture, when one CPU runs the irq17: bce1 aacu0 thread and another one runs arcconf. I wonder if that might be a source of bad locking or races, or.. The arcconf utility uses ioctl that goes into aac/aacu(4) internals. Do you see the same result w/ the in-tree aac(4) driver as opposed to Adaptec's version? -Ed [It's quite hard to move back to aac(4) as that requires fstab update [ aacdu0 - aacd0] and instant reboot, because we use quotas and quotacheck looks into /etc/fstab. Such preparations as fstab update and commenting out load_aacu=YES will give discrepancy between fstab and actual mount points.] I will try anyway. Thank you for your help. -- wbr, pluknet ___ 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: 6.2 sporadically locks up
2009/6/19 Adrian Chadd adr...@freebsd.org: Just modify the driver slightly to hijack a different device prefix :) Hi, Adrian. That's where I just go if I should have to. - .d_name = aac, + .d_name = aacu, (or vise versa) While here, I'd like to give some summary about locking up with irq17:bce1 aacu0 vs arcconf scenario. Abstract: we have a number of boxes with IBM ServeRAID 8k on 6.2. That scenario takes place only with aacu b15753, and not with b15411 (at least not noticed). We take a decision some time ago to move some boxes to 6.4 (and leave vendor aacu b15753 there as it's) to see how it goes. Until now (2 or 3 weeks) there were no lockup. I hope it will so farther.. Adrian 2009/6/17 pluknet pluk...@gmail.com: 2009/6/17 Ed Maste ema...@freebsd.org: On Tue, Jun 16, 2009 at 07:03:34PM +0400, pluknet wrote: As for allpcpu, I often see the picture, when one CPU runs the irq17: bce1 aacu0 thread and another one runs arcconf. I wonder if that might be a source of bad locking or races, or.. The arcconf utility uses ioctl that goes into aac/aacu(4) internals. Do you see the same result w/ the in-tree aac(4) driver as opposed to Adaptec's version? -Ed [It's quite hard to move back to aac(4) as that requires fstab update [ aacdu0 - aacd0] and instant reboot, because we use quotas and quotacheck looks into /etc/fstab. Such preparations as fstab update and commenting out load_aacu=YES will give discrepancy between fstab and actual mount points.] I will try anyway. Thank you for your help. -- wbr, pluknet ___ 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