/var full and fails fsck -y

[ecorban]

Hi
i get the following error message from fsck -y when i try to fsck -y /dev/vg00/lvol8.

I tried to extend /dev/vg00/lvol8 >> /var
but could not through the console port as single user mode. is there any way to fix this error ?
i'll give points for fix the problem.....

mountall: /dev/vg00/lvol8 has to be fsck'd
mountall: diagnostics from fsck
vxfs fsck: sanity check: /dev/vg00/lvol8 needs checking
mountall: cannot fsck /dev/vg00/lvol8
mountall: diagnostics from fsck
vxfs fsck: cannot initialize aggrefile system check failure, aborting ...
mountall: /dev/vg00/lvol8 failed in fsck
mountall: /dev/vg00/lvol8 cannot be mounted

[John Poff]

Hi,

If this is a VxFS filesystem, you can try the fsck with the full log option:

fsck -F vxfs -o full /dev/vg00/rlvol8

Plus, I would try it with the raw logical volume (rlvol8). If this fails, do an 'echo $?' to see what error code is returned.

JP

[Sundar_7]

I dont believe you could do a successful fsck when there is no free space in the filesystem

Not sure if fsadm/extendfs is going to work either. I know if fsadm is unable to create the lock file,it will fail.

How did you try to extend the filesystem ? -what was the error message ?

[ecorban]

I used extendfs as single user. then after a couple of days usage the fs filled with some panic. dont know why that. now its full and i cant fix it.

The error no. i got was an error number 16.

anyway im connected to the console from a HP term and all the ascii characters are all screwed up. its doesnt page correctly. every letter and number is scrolled though the one line.

anyhelp here with this nonsense is helpful. i tried settingthe server to defaults on the ISL area but still doesnt help

[ecorban]

also when i boot the host, the host creates a /var but it doesnt show under the bdf. it contains just some reuimentary directories. but any stuff like sam or ignite is not visible.............

[John Poff]

Nothing is showing under /var because the filesystem isn't getting mounted.

For your console, try getting into the settings and see what term type it is set for. I'm not a wizard with those console settings, but I seem to remember that some of the newer systems expect the console to be VT100 instead of HPTERM or HP70090.

JP

[John Poff]

Here is a thread that discusses the term settings on a console and how to fix them:

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=26943

JP

[ecorban]

sorted out the console settings.

my problem is this:
one ofthe disks assigned to my /dev/vg00 has failed and doesnt respond to any probing (diskinfo etc). i attached a new disk to the host but it fails to vgextend the /dev/vg00 group as the failed disk doesnt respond and vgextend justs bombs out entirely. how do i recover the system ?

[Steven E. Protter]

Quick recovery:

If you have a make_tape_recovery tape, you can replace the bad disk, boot off the tape and restore vg00 to where it was before the failure.

You need to pvcreate the new disk before you can use it.

ioscan -fnCk disk
ioscan -fnC disk
or
ioscan

Find out if it shows up at all.

Obtain the physical address of the new drive.

Lets day its /dev/dsk/c1t6d0

pvcreate /dev/dsk/c1t6d0
vgextend

onward and upward.

[ecorban]

no tape backup.. i wish i had........