FreeBSD bugmaster | 5 Nov 12:06 2007
Picon

Current problem reports assigned to freebsd-fs <at> FreeBSD.org

Current FreeBSD problem reports
Critical problems
Serious problems

S Tracker      Resp.      Description
--------------------------------------------------------------------------------
o kern/112658  fs         [smbfs] [patch] smbfs and caching problems (resolves b
o kern/114676  fs         [ufs] snapshot creation panics: snapacct_ufs2: bad blo
o kern/114856  fs         [ntfs] [patch] Bug in NTFS allows bogus file modes.
o kern/116170  fs         Kernel panic when mounting /tmp

4 problems total.

Non-critical problems

S Tracker      Resp.      Description
--------------------------------------------------------------------------------
o kern/114847  fs         [ntfs] [patch] dirmask support for NTFS ala MSDOSFS

1 problem total.

_______________________________________________
freebsd-fs <at> freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-fs
To unsubscribe, send any mail to "freebsd-fs-unsubscribe <at> freebsd.org"

Andrew Thompson | 5 Nov 22:50 2007
Picon

Re: ZFS kmem_map too small.

On Mon, Oct 08, 2007 at 02:15:23PM +0200, Pawel Jakub Dawidek wrote:
> Here are some updates:
> 
> I was able to reproduce the panic by rsyncing big files and trying
> bonnie++ test suggested in this thread.
> 
> Can you guys retry with this patch:
> 
> 	http://people.freebsd.org/~pjd/patches/vm_kern.c.2.patch
> 
> It's a hack, yes, but allows to mitigate the problem quite well. I'm
> looking for a solution that can be used for 7.0 before we find a better
> fix.

Do you still plan to commit this for 7.0 or is a better solution in the
works? I would assume something would need to be in before RC1.

> BTW. To use ZFS you _must_ increase vm.kmem_size/vm.kmem_size_max.
> If you have the problem discussed here and you're using standard values,
> please retry with vm.kmem_size/vm.kmem_size_max set to at least 600MB in
> /boot/loader.conf.
> 
> I'm not sure if it's not too late to ask re <at>  about increasing the
> default kmem size at least on amd64. ~300MB we have there is silly
> small.
> 
> -- 
> Pawel Jakub Dawidek                       http://www.wheel.pl
> pjd <at> FreeBSD.org                           http://www.FreeBSD.org
> FreeBSD committer                         Am I Evil? Yes, I Am!
(Continue reading)

Niki Denev | 6 Nov 00:31 2007

Re: ZFS kmem_map too small.

On Nov 5, 2007 11:50 PM, Andrew Thompson <thompsa <at> freebsd.org> wrote:
> On Mon, Oct 08, 2007 at 02:15:23PM +0200, Pawel Jakub Dawidek wrote:
> > Here are some updates:
> >
> > I was able to reproduce the panic by rsyncing big files and trying
> > bonnie++ test suggested in this thread.
> >
> > Can you guys retry with this patch:
> >
> >       http://people.freebsd.org/~pjd/patches/vm_kern.c.2.patch
> >
> > It's a hack, yes, but allows to mitigate the problem quite well. I'm
> > looking for a solution that can be used for 7.0 before we find a better
> > fix.
>
> Do you still plan to commit this for 7.0 or is a better solution in the
> works? I would assume something would need to be in before RC1.
>
> > BTW. To use ZFS you _must_ increase vm.kmem_size/vm.kmem_size_max.
> > If you have the problem discussed here and you're using standard values,
> > please retry with vm.kmem_size/vm.kmem_size_max set to at least 600MB in
> > /boot/loader.conf.
> >
> > I'm not sure if it's not too late to ask re <at>  about increasing the
> > default kmem size at least on amd64. ~300MB we have there is silly
> > small.
> >

I have fairly easy reproductable (bonnie++) kmem_map too small panics
on AMD64 even with kmem_size and kmem_size_max to 1G on 8Gig machine.
(Continue reading)

Borja Marcos | 6 Nov 10:04 2007
Picon

Re: ZFS kmem_map too small.


On Nov 6, 2007, at 12:31 AM, Niki Denev wrote:
> I have fairly easy reproductable (bonnie++) kmem_map too small panics
> on AMD64 even with kmem_size and kmem_size_max to 1G on 8Gig machine.
> I've tried setting them to 2G but the kernel panics on boot.
> I have compiled in debugger support and can provide backtraces and
> testing if needed (the machine is still not in production).

Same here. I can provoke an instant panic running (for example) a make  
buildworld and
bonnie++. I have all the filesystems on ZFS, booting from a USB  
pendrive. I tried to set
the kernel memory limits to 1.5 GB but it still crashes with the  
kmem_too_small panic. I also
tried the patch supplied by Pawel.

The machine has 6 hard disks configured as a raidz2 pool.

Please let me know of any test anyone might need.

Borja.

----------------
"The thing he realised about the windows was this: because they had  
been converted into openable windows after they had first been  
designed to be impregnable, they were, in fact, much less secure than  
if they had been designed as openable windows in the first place."
    Douglas Adams, "Mostly Harmless"

_______________________________________________
(Continue reading)

Pawel Jakub Dawidek | 6 Nov 11:00 2007
Picon

Re: ZFS kmem_map too small.

On Tue, Nov 06, 2007 at 10:04:43AM +0100, Borja Marcos wrote:
> 
> On Nov 6, 2007, at 12:31 AM, Niki Denev wrote:
> >I have fairly easy reproductable (bonnie++) kmem_map too small panics
> >on AMD64 even with kmem_size and kmem_size_max to 1G on 8Gig machine.
> >I've tried setting them to 2G but the kernel panics on boot.
> >I have compiled in debugger support and can provide backtraces and
> >testing if needed (the machine is still not in production).
> 
> Same here. I can provoke an instant panic running (for example) a make  
> buildworld and
> bonnie++. I have all the filesystems on ZFS, booting from a USB  
> pendrive. I tried to set
> the kernel memory limits to 1.5 GB but it still crashes with the  
> kmem_too_small panic. I also
> tried the patch supplied by Pawel.
> 
> The machine has 6 hard disks configured as a raidz2 pool.
> 
> Please let me know of any test anyone might need.

If you use vm_kern.c.2.patch, can you show loader.conf and exact command
that can provke the panic?

--

-- 
Pawel Jakub Dawidek                       http://www.wheel.pl
pjd <at> FreeBSD.org                           http://www.FreeBSD.org
FreeBSD committer                         Am I Evil? Yes, I Am!
Borja Marcos | 6 Nov 11:20 2007
Picon

Re: ZFS kmem_map too small.


On Nov 6, 2007, at 11:00 AM, Pawel Jakub Dawidek wrote:

>> The machine has 6 hard disks configured as a raidz2 pool.
>>
>> Please let me know of any test anyone might need.
>
> If you use vm_kern.c.2.patch, can you show loader.conf and exact  
> command
> that can provke the panic?

loader.conf:
zfs_load="YES"
vfs.root.mountfrom="zfs:pool/root"
vm.kmem_size_max="1610612736"
vm.kmem_size="1610612736"

Y bumped up the vnodes as well, as suggested at the ZFS tuning Wiki.

(/etc/sysctl.conf)
kern.maxvnodes: 400000

Anything else? (I just don't want to spam a big message with kernel  
config, etc)

I don't have a fancy config as far as I know. I'm running the latest  
FreeBSD 7-BETA, amd64,
and the machine has 4 GB of memory. It's a Dell Poweredge 2950, I'm  
using the mfi passthough
devices as disks (I don't want to use the disk array functionality it  
(Continue reading)

Borja Marcos | 6 Nov 11:38 2007
Picon

Re: ZFS kmem_map too small.


On Nov 6, 2007, at 11:00 AM, Pawel Jakub Dawidek wrote:

> If you use vm_kern.c.2.patch, can you show loader.conf and exact  
> command
> that can provke the panic?

Sorry, forgot about the command.

I just created a ZFS volume (scratch) and used:

bonnie++ -s 12g -n 24

The machine dies consistently during the single file tests (sequential  
write, rewrite, etc).

Borja.

----------------
"The thing he realised about the windows was this: because they had  
been converted into openable windows after they had first been  
designed to be impregnable, they were, in fact, much less secure than  
if they had been designed as openable windows in the first place."
    Douglas Adams, "Mostly Harmless"

_______________________________________________
freebsd-fs <at> freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-fs
To unsubscribe, send any mail to "freebsd-fs-unsubscribe <at> freebsd.org"

(Continue reading)

Ivan Voras | 6 Nov 12:11 2007
Picon

Re: ZFS kmem_map too small.

Borja Marcos wrote:

> 
> Y bumped up the vnodes as well, as suggested at the ZFS tuning Wiki.
> 
> (/etc/sysctl.conf)
> kern.maxvnodes: 400000

This particular setting was suggested to increase performance, not
reliability - I'll clarify that on the page.

_______________________________________________
freebsd-fs <at> freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-fs
To unsubscribe, send any mail to "freebsd-fs-unsubscribe <at> freebsd.org"

Pawel Jakub Dawidek | 6 Nov 13:55 2007
Picon

Re: ZFS kmem_map too small.

On Tue, Nov 06, 2007 at 11:20:34AM +0100, Borja Marcos wrote:
> 
> On Nov 6, 2007, at 11:00 AM, Pawel Jakub Dawidek wrote:
> 
> >>The machine has 6 hard disks configured as a raidz2 pool.
> >>
> >>Please let me know of any test anyone might need.
> >
> >If you use vm_kern.c.2.patch, can you show loader.conf and exact  
> >command
> >that can provke the panic?
> 
> loader.conf:
> zfs_load="YES"
> vfs.root.mountfrom="zfs:pool/root"
> vm.kmem_size_max="1610612736"
> vm.kmem_size="1610612736"
> 
> Y bumped up the vnodes as well, as suggested at the ZFS tuning Wiki.
> 
> (/etc/sysctl.conf)
> kern.maxvnodes: 400000

The Wiki should be changed. Allow ZFS to autotune it, don't tune it by
hand.

> Anything else? (I just don't want to spam a big message with kernel  
> config, etc)

I asked about exact command to trigger the panic.
(Continue reading)

Pawel Jakub Dawidek | 6 Nov 13:55 2007
Picon

Re: ZFS kmem_map too small.

On Tue, Nov 06, 2007 at 11:38:55AM +0100, Borja Marcos wrote:
> 
> On Nov 6, 2007, at 11:00 AM, Pawel Jakub Dawidek wrote:
> 
> >If you use vm_kern.c.2.patch, can you show loader.conf and exact  
> >command
> >that can provke the panic?
> 
> Sorry, forgot about the command.
> 
> I just created a ZFS volume (scratch) and used:

ZFS volume? You mean ZVOL? Can you paste 'zfs list; zpool status'?

> bonnie++ -s 12g -n 24
> 
> The machine dies consistently during the single file tests (sequential  
> write, rewrite, etc).

--

-- 
Pawel Jakub Dawidek                       http://www.wheel.pl
pjd <at> FreeBSD.org                           http://www.FreeBSD.org
FreeBSD committer                         Am I Evil? Yes, I Am!

Gmane