Autoreboot - no Dump Device & Autofsck

T

thavinci

Hi.

Because of zfs i find servers crashing on a regular basis, however the biggest issue is that they never come back online tilll i manually push reset next day and fsck the damaged partitions.

On a kernel panic it sits at the reboot in 15sec screen and never does.
There is a complain about

Code: 
Cannot dump. Device not defined or unavailable

now my sysctl.conf has
Code: 
kern.coredump=0
kern.corefile=/dev/null

Specified and im wondering it thats what's preventing reboot.

Also as mentioned on reboot the server will exit to single user mode and expect me to run fsck manually.

Is there now way it can just go
Code: 
fsck -y "filesystem
itself?
 
remove
Code: 
kern.corefile=/dev/null
from /etc/sysctl.conf

so you say, that you use zfs.... and that you need to run fsck manyally?
Zfs deosn't need fsck... in fact, there is no fsck for zfs :)

tell us about your partition layout
and show output of $ mount
 
I had no
Code: 
kern.corefile=/dev/null
specified and same result.

Reason for fsck is because all boot drives are still ufs i wouldn't want more issues by making it zfs too:

Code: 
 mount                                                        Press ? for help
/dev/ad4s1a on / (ufs, local, noatime)
devfs on /dev (devfs, local, multilabel)
TwoTB1 on /mnt/2tb1 (zfs, local, noatime)
TwoTB2 on /mnt/2tb2 (zfs, local, noatime)
tank/Nico on /mnt/Nico (zfs, local, noatime)
tank/Sam on /mnt/Sam (zfs, local, noatime)
tank/lptdrv on /mnt/lptdrv (zfs, local, noatime)
tank on /mnt/tank (zfs, local, noatime)
tank/Backups on /mnt/tank/Backups (zfs, NFS exported, local, noatime)
tank/Docs on /mnt/tank/Docs (zfs, local, noatime)
devfs on /var/named/dev (devfs, local, multilabel)
/mnt/tank/Storage/Series on /mnt/ftp/Series (nullfs, local)
/mnt/tank/Storage/Mp3's on /mnt/ftp/Mp3s (nullfs, local)
/mnt/2tb2/Movies on /mnt/ftp/Movies (nullfs, local)
 
[cmd=]$ grep fsck /etc/defaults/rc.conf[/cmd] should give you an idea.
 
Much appreciated i went with:

Code: 
fsck_y_enable="YES"
background_fsck="NO"

Solves one issue. The server crashed again last night and this seems to have done trick with the fsck part. Now too get it to autoreboot on panic. :>
 
Code: 
$ sysctl -ad |grep dump
kern.coredump: Enable/Disable coredumps

Set kern.coredump="0" in /etc/sysctl.conf to disable kernel dumps on panic.

From /usr/src/sys/conf/NOTES:
Code: 
# Set the amount of time (in seconds) the system will wait before
# rebooting automatically when a kernel panic occurs.  If set to (-1),
# the system will wait indefinitely until a key is pressed on the
# console.
options         PANIC_REBOOT_WAIT_TIME=16

Which should already be set, but you can try setting it to 5 or something low like that (it's in seconds).

You'll need to recompile the kernel, obviously. :)
 
Thing is that i already have
Code: 
kern.coredump=0
kern.corefile=/dev/null
specified so it shoulnd't be trying to dump or so says my logic.

Yet it still tries too and what's worse i don't know if it's because of the
Code: 
Cannot dump. Device not defined or unavailable
message that it simply gets stuck at
Code: 
Autoimatic reboot in 15 seconds - press a key on the console to abort

Or something else, as i cannot press a key either.

Looks like so:



Uploaded with ImageShack.us
 
You must log in or register to reply here.
Back
Top