kernel panic

W

wonslung

hello, recently got my zfs server up and running.

running Freebsd 7.2-stable amd 64 intel q9550 8gb ddr2 800

i was doing a cp and i receieved this error:
Code: 
panic: kmem_malloc (114688): kmem_map too small: 2697601024 total allocated cpuid = 2

I think i got it right, i wrote it down because i didn't know where to find it later.\


reading "UPDATING", does this mean i need to set vm.kmem_size_max to something else?

the message in UPDATING is as follows
20090207:
ZFS users on amd64 machines with 4GB or more of RAM should
reevaluate their need for setting vm.kmem_size_max and
vm.kmem_size manually. In fact, after recent changes to the
kernel, the default value of vm.kmem_size is larger than the
suggested manual setting in most ZFS/FreeBSD tuning guides.

if so what would be a good setting for 8 gb ram?

edit:
This has happened every single time i try to do a LARGE copy with copy. It doesn't happen at the same point...what i've noticed is that when using cp and watching zpool iostat -v when disk i/o gets to 200-250 MB/s it panics and crashes....using piped tars to copy stuff isn't causing the same panic or nearly as high load.

I'm really unsure what's happening...if anyone knows more please help, everything i've read on google says amd64 shouldn't need any tuning for ZFS but obviously i do.
 
With 64-bit FreeBSD 7.2 and later, you don't need to tune kmem_* settings in /boot/loader.conf. Nor are you forced to tune zfs.arc_size_max.

Comment out, or remove, those lines from /boot/loader.conf, and see if it still panics during the copy.
 
lol, i DIDNT use them

thats my point.
I got the kernel panic and i'm wondering WHY

i thought that i DIDNT have to tune it....but when i do a command like cp every so often it panics and crashes with the above message, has happened 3 times in 2 days now
 
Oh. Sorry. :)

In that case, you can try to tune it. :)

Try setting just the ARC size to something like 2G and see how that goes.
 
remember when i asked about spreading the stuff across all the vdevs because i had to add one later?

to acomplish this i first copied all the data to ZFS from the single drives to the 2 raidzvdev system.

after that, i added the 3rd raidz vdev and copied the data to a new zfs filesystem in the pool and deleted it afterwards

what i noticed was when i used the cp command for this, it would panic

when i used cpio or piped tar, it wouldn't panic...also cp seemed to stress the cpu a lot more...

now, i have 8gb of ram and it MIGHT have been because i'm using the zvol as swap...what do you think?


also, i added a log device, an 80gb hard drive, i'm wondering now if i can safely remove it.

my thoughts are i could slice it because the ZIL never uses much, 100-200 mb tops so far

i could put the swap on THAT if i can safely remove the zil
 
You must log in or register to reply here.
Back
Top