Vg01 cannot mount on hp9000

[Nsikan_1]

Hi all,
I have an hP9000 server running hpux 10.20.The problem is that client said there was a power failure while restoring a backup,after which vg01 residing on model 10 disk array cannot be mounted.But the VG00 is mounted and OS running perfectly okay.I checked the file systems from /etc/fstab , and discovered that the file system on vg01 are hfs.I tried all I know to get vg01 mounted ,but could not.
errors displayed include:
vg01 cannot be activated if tried to activate.
"powerfail" error will be displayed.
the device file: /dev/dsk/c1t0d0 does not have LVM configuration,this the path of the disk array.
Does anybody have such experience?Pls i need ur assitant as yesterday.
rgds.
Nsikan

P.S. This thread has been moved from Servers > HP 9000 to HP-UX > LVM - HP Forums Moderator

[Ricardo Rocha]

you could have a disk error. how many disks are in vg01 and are the lv mirrored?

[MANOJ SRIVASTAVA]

It look as if the disk ahs lost the lvm config , try doing a vgcfgrestore ? or in case you can call HP for support .


the syntax is


vgcfgrestore -n /dev/vg01 /dev/rdsk/c1t0d0 in case the disk is fine .


then do vgchange -a y /dev/vg01

and fsck if the lvols dont come up.


Manoj Srivastava

[S.K. Chan]

It looks like vg01 might not have its quorum requirement, thus "vgchange -a y vg01" (ie activating it) failed. Take a look at /var/adm/syslog/syslog.log file again and see if there are any other detail error. What does ..
# diskinfo /dev/rdsk/
return ?
I'm suspecting some connection problem with your SCSI cable and/or termination. It may be also one or more disks in vg01 are down.

[Nsikan_1]

yeah,
I have 5 disk on the array. it is a model 10 disk array.the scsi cable is good b/c, ioscan can see the disk.if i run diskinfo it gives the correct info on it.
I thought of exporting the Vg01 an import again, but it didn't work.It export but cannot import.
more research pls.
Nsikan.

[S.K. Chan · Accepted solution]

Are you saying vg01 is already exported ? And you have problem importing it back ? Check ..
# ll /dev/vg01
and
# strings /etc/lvmtab
If vg01 has been exported its device files should not exist, so does the enty in "lvmtab" file.
When you exported it did you specify a "mapfile" ? If so import it back like this ..

# mkdir /dev/vg01
# mknod /dev/vg01/group c 64 0x010000
==> Minor number 0x010000 must be unique.
# vgimport -m /mapfile /dev/vg01 /dev/dsk/ /dev/dsk/ ...
==> Assuming you have the mapfile in / after you did the vgexport. If you do not have the map file , vgimport it without "-m /mapfile" option.
# vgchange -a y /dev/vg01
==> Activate it.
# vgcfgbackup /dev/vg01
==> Back up LVM config to /etc/lvmconf.
# mount -a
==> Mount all FS back.

[Nsikan_1]

Hi Chan,
I did all the steps u mentioned ,but got stock at vgimport stage,i did not use the mapfile, but it shd have gone even.
its given me those errors I mentioned.
More replies pls.
Nsikan

[Ricardo Rocha]

hi again
try typing
dd if=/dev/rdsk/c1t0d0 of=/dev/null bs=512k
in order to identify physically the disk (it will blink)
if you have a mirrored structure lvm (but only if you have it) you may try to activate the vg01 with this disk out of the sc10 - the disk may be filling the bus with garbage....
if it works you should replace the disk

[Ricardo Rocha]

ok
just noticed you have 5 disks, so no mirroring!
can you activate the vg01 with the no quorum option? (-q n)

[Nsikan_1]

hi all,
pls note that:
a. when restarting previous backup,the vg & fs are mounted respectively,but when shutdown to restart,it can't be mounted.

b. It's using module 10 disk array - Nike/Galaxy with 5disks bined as vg01

Also:
1. If running vgcfgrestore, it takes more than necessary without result.ps -ef |grep vgcfgrestore shows nothing going on.
2. after following the steps to import the disk accordingly, ll /dev/vg01 display accordingly
3. strings /etc/lvmtab shows
a. /dev/vg00 & its device accordingly
b. after the export, vg01 not showing.
4. /etc/fstab, shows all the files accordingly but with "hfs" type b/4 editing fstab.
5. strings /etc/mnttab shows only vg00 & its mounted fs.
6.vgscan: couldn't query pv "/dev/dsk/c1t0d0" , the specified path does not correspond tp pv attached to the vg. warning: couldn't query all of the pvs,unable to query th i/o interface: parameter error unable to scan pv on system
7. vgsync vg01:
-couldn't query "/dev/vg01/lvol2": vg not activated.
-couldn't query the list of lvs.
-couldn't access th list of lvs for vg -vg01
couldn't resynchronize vg "vg01"
8. vgchange -a y vg01
-couldn't attach to the vg pv "/dev/dsk/c1t0d0"
The path of the pv refers to a device that does not exist, or is not configured into the kernel.
vgchange:warning:couldn't query pv "/dev/dsk/c1t0d0"
the specified path does not correspond to pv attached to this vg.
vgchange: couldn't activate vg01
quorum not present or some pv are missing.

n/b: but "ioscan" shows /dev/dsk/c1t0d0 as the path to diskarray- module 10.

9. vgimport -v /dev/dsk/c1t0d0 shows this:
Begining the import process on vg "/dev/dsk/c1t0d0"
vgimport: cannot open the contro file "/dev/dsk/c1t0d0"
Not a directory.

I'm thinking of perhaps deleting the lun completely & recereate it, but won't do it,what are the steps,it has no mux port ,cables connected directly from system port to the array ports accordingly.
I did it b/4 but was using the one with a mux.

more resaerch pls.the client is becoming impatient.

Nsikan.