TruCluster
Showing results for 
Search instead for 
Do you mean 

Tru64 5.1b PK3 / showfdmn error

Occasional Advisor

Tru64 5.1b PK3 / showfdmn error

Hello

we have a 2 Node GS1280 Cluster with Tru64 5.1b PK3 and 4 Patches:

Patches installed on the system came from following patch kits:
--------------------------------------------------------------

- T64KIT0020747-V51BB24-20031118 OSF540
- T64KIT0020861-V51BB24-E-20031125 OSF540
- T64KIT0021750-V51BB24-20040227 OSF540
- T64KIT0021909-V51BB24-E-20040317 OSF540
- T64V51BB1AS0001-20021229 OSF540
- T64V51BB1AS0001-20021229 TCR540
- T64V51BB24AS0003-20030929 OSF540
- T64V51BB24AS0003-20030929 TCR540

the Problem is, when I try the Command:

root@tru64:/var/adm# showfdmn root6_domain
showfdmn: unable to get info for domain 'root6_domain'
showfdmn: error = Device busy
root@tru64:/var/adm#

I get this error, that men's that we can't use the command showfdmn on root6_domain.

In the bootprocess the following error appears:

.
.
Sep 18 09:57:42 liux060f vmunix: CMS: Joining deferred filesystem sets
Sep 18 09:57:42 liux060f vmunix: msfs_mount: The mount device does not match the linked device.
Sep 18 09:57:42 liux060f vmunix: Check linked device in /etc/fdmns/domain
Sep 18 09:57:42 liux060f vmunix: msfs_mount: Setting boot device name to local_root RW
Sep 18 09:57:42 liux060f vmunix: CNX QDISK: Successfully claimed quorum disk, adding 1 vote.
.
.



everthing else works fine.

Has someone any Idea?

thanks for help mike
10 REPLIES
Honored Contributor

Re: Tru64 5.1b PK3 / showfdmn error

No solution, just soem places where I would look.

First... re-evaluate the LSM data.
Is LSM entirely happy?

Next "Device busy" generally means trying an operation on a device that should not have been mounted. But showfdmns works on mounted devices.
So why the confusion? Carefully check the structures under /etc/fdmns/

May a metadata corruption? Try fixfdmn (in read only mode first)?

fwiw,
Hein.
Honored Contributor

Re: Tru64 5.1b PK3 / showfdmn error

If the /etc/fdmns directory structure is corrupted, you can try reconstructing it with '/sbin/advfs/advscan' command.

Honored Contributor

Re: Tru64 5.1b PK3 / showfdmn error

Hi,

can you please post ls -lR /etc/fdmns
?
I wonder if a link points to a wrong device.

greetings,

Michael
Occasional Advisor

Re: Tru64 5.1b PK3 / showfdmn error

Morning

great forum here, I was very surprised about the quick responses.

thx all

here the output from "ls -lR /etc/fdmns":


root@liux060f:/# ls -lR /etc/fdmns
total 120
-r-------- 1 root system 0 Oct 7 12:14 .advfslock_alt_boot5
-r-------- 1 root system 0 Oct 7 12:26 .advfslock_alt_cluster_root
-r-------- 1 root system 0 Oct 7 12:29 .advfslock_alt_cluster_usr
-r-------- 1 root system 0 Oct 7 12:14 .advfslock_alt_cluster_var
-r-------- 1 root system 0 Feb 21 2002 .advfslock_cluster_root
-r-------- 1 root system 0 Feb 21 2002 .advfslock_cluster_usr
-r-------- 1 root system 0 Feb 21 2002 .advfslock_cluster_var
-r-------- 1 root system 0 Jul 5 2002 .advfslock_docdom1
-r-------- 1 root system 0 Feb 20 2002 .advfslock_fdmns
-r-------- 1 root system 0 Jun 20 08:17 .advfslock_p02archdom1
-r-------- 1 root system 0 Jun 20 08:13 .advfslock_p02datadom1
-r-------- 1 root system 0 Jun 20 10:40 .advfslock_p02datadom2
-r-------- 1 root system 0 Jun 20 08:31 .advfslock_p02datadom3
-r-------- 1 root system 0 Jun 20 10:24 .advfslock_p02datadom4
-r-------- 1 root system 0 Jun 20 07:53 .advfslock_p02datalog1
-r-------- 1 root system 0 Jun 20 08:08 .advfslock_p02datalog2
-r-------- 1 root system 0 Jun 20 10:42 .advfslock_p02userdom1
-r-------- 1 root system 0 Jul 17 13:34 .advfslock_root5_domain
-r-------- 1 root system 0 Aug 2 12:12 .advfslock_root6_domain
-r-------- 1 root system 0 Mar 7 2003 .advfslock_scratchdom1
drwxr-xr-x 2 root system 8192 Jun 20 17:02 cluster_root
drwxr-xr-x 2 root system 8192 Jun 20 17:02 cluster_usr
drwxr-xr-x 2 root system 8192 Jun 20 17:02 cluster_var
drwxr-xr-x 2 root system 8192 Jun 20 16:54 docdom1
drwxr-xr-x 2 root system 8192 Jun 20 16:54 p02archdom1
drwxr-xr-x 2 root system 8192 Jun 20 16:54 p02datadom1
drwxr-xr-x 2 root system 8192 Jun 29 04:15 p02datadom2
drwxr-xr-x 2 root system 8192 Jun 20 16:54 p02datadom3
drwxr-xr-x 2 root system 8192 Jun 20 16:54 p02datadom4
drwxr-xr-x 2 root system 8192 Jun 20 16:54 p02datalog1
drwxr-xr-x 2 root system 8192 Jun 20 16:54 p02datalog2
drwxr-xr-x 2 root system 8192 Jun 20 16:54 p02userdom1
drwxr-xr-x 2 root system 8192 Jul 17 13:34 root5_domain
drwxr-xr-x 2 root system 8192 Aug 2 12:12 root6_domain
drwxr-xr-x 2 root system 8192 Jun 20 16:54 scratchdom1

/etc/fdmns/cluster_root:
total 0
lrwxr-xr-x 1 root system 15 Jun 20 17:02 dsk3a -> /dev/disk/dsk3a

/etc/fdmns/cluster_usr:
total 0
lrwxr-xr-x 1 root system 15 Jun 20 17:02 dsk3d -> /dev/disk/dsk3d

/etc/fdmns/cluster_var:
total 0
lrwxr-xr-x 1 root system 15 Jun 20 17:02 dsk3e -> /dev/disk/dsk3e

/etc/fdmns/docdom1:
total 0
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk58c -> /dev/disk/dsk58c

/etc/fdmns/p02archdom1:
total 0
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk55c -> /dev/disk/dsk55c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk56c -> /dev/disk/dsk56c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk57c -> /dev/disk/dsk57c

/etc/fdmns/p02datadom1:
total 0
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk35c -> /dev/disk/dsk35c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk36c -> /dev/disk/dsk36c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk37c -> /dev/disk/dsk37c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk38c -> /dev/disk/dsk38c

/etc/fdmns/p02datadom2:
total 0
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk39c -> /dev/disk/dsk39c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk40c -> /dev/disk/dsk40c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk41c -> /dev/disk/dsk41c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk42c -> /dev/disk/dsk42c
lrwxr-xr-x 1 root system 16 Jun 29 04:15 dsk68c -> /dev/disk/dsk68c

/etc/fdmns/p02datadom3:
total 0
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk43c -> /dev/disk/dsk43c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk44c -> /dev/disk/dsk44c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk45c -> /dev/disk/dsk45c

/etc/fdmns/p02datadom4:
total 0
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk46c -> /dev/disk/dsk46c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk47c -> /dev/disk/dsk47c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk48c -> /dev/disk/dsk48c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk49c -> /dev/disk/dsk49c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk50c -> /dev/disk/dsk50c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk51c -> /dev/disk/dsk51c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk52c -> /dev/disk/dsk52c

/etc/fdmns/p02datalog1:
total 0
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk59c -> /dev/disk/dsk59c

/etc/fdmns/p02datalog2:
total 0
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk60c -> /dev/disk/dsk60c

/etc/fdmns/p02userdom1:
total 0
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk53c -> /dev/disk/dsk53c
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk54c -> /dev/disk/dsk54c

/etc/fdmns/root5_domain:
total 0
lrwxr-xr-x 1 root system 16 Jul 17 13:34 dsk82a -> /dev/disk/dsk82a

/etc/fdmns/root6_domain:
total 0
lrwxr-xr-x 1 root system 16 Aug 2 12:12 dsk32a -> /dev/disk/dsk32a

/etc/fdmns/scratchdom1:
total 0
lrwxrwxrwx 1 root system 16 Jun 20 16:54 dsk61c -> /dev/disk/dsk61c

the Cluster is fine, all checks are fine.
( cfsmgr -v, clu_check_config, clu_get_info -full )

The Problem is that two of our Scripts use this command, and we can't use this Scripts anymore. And also we think at the next update/upgrade we become serious Problems.


cheers

Mike
Honored Contributor

Re: Tru64 5.1b PK3 / showfdmn error

have a look to:

Sep 18 09:57:42 liux060f vmunix: msfs_mount: The mount device does not match the linked device.

So it seems to be an error in configuration, does all members have the same problem or only one member?

If it is only one member correct the links or delete/recreate cluster member which is the fastest way to solve the problem. Another question is if you have changed something like bootdevice, etc. or swapped hardware without correcting configuration.
Help() { FirstReadManual(urgently); Go_to_it;; }
Esteemed Contributor

Re: Tru64 5.1b PK3 / showfdmn error

Could you try relocating the CFS server for that domain to another node, e.f. "cfsmgr -a SERVER=anothernode -d root6_domain"?
Occasional Advisor

Re: Tru64 5.1b PK3 / showfdmn error

Hi Ralf

No only Member Nr. 6 have the Problem. We already create thrice this member, and it doesn't solve the Problem. We change nothing the Problem happens from the beginning on.

Cluster History:

First we had two ES45 ( Member 1 and 2 ) than we migrate to ES80 ( Member 3 and 4 ) with HSG80 Storage. Than we migrate to XP Storage ( we have 3 Clustermemberdisks, 2 for the aktual Members and one for Migration )

Then we add one new GS1280 to the Cluster ( Member 5 ) and remove Member 3 from the Cluster. Then we take the free memberdisk from Member 3 and create the Member 6 from Scratch and boot into the Cluster. All goes fine but when i check the bootlog I see the error and now we try to find the Problem and how can we solv the Problem.

cheers mike
Honored Contributor

Re: Tru64 5.1b PK3 / showfdmn error

do not delete and recreate the cluster member with the same id, use another instead.

Be sure this member do have valid pathes to the boot devices. Error messages sound like a device problem or is a follow on to previous action. It is similar to problems resulting from "cloning" or "booting" disk from other members which is unsupported.

So best try is to create an additional cluster member on a new disk for this member, boot it and check if problem persist. If not delete old member configuration.

Help() { FirstReadManual(urgently); Go_to_it;; }
Honored Contributor

Re: Tru64 5.1b PK3 / showfdmn error


Could it be that you are booted from a valid bootdisk but that it is not really dsk32 ?
Check the disklabel of dsk32 and see what memberid it gives in the volume field. Also verify it's h-partition speciale file major/minor number and check if they agree with the seqdisk_major/seqdisk_minor in the clubase section of this members sysconfigtab.

Also check the output of clu_bdmgr -d dsk32

_Johan_

_JB_
Occasional Advisor

Re: Tru64 5.1b PK3 / showfdmn error

Hi all

thanks for all replys. The Problem is solved!
The Solution was: We don't now ! After the weekend it works perfekt, no problem. Now I wait for the next Maintenace Window to reboot the Clustermember and lock whats happen. We hope that now all is fine.

On Monday i try again the "showfdmn root6_domain" command, because the HP - Engineer told me. And it works, and I swear we don't change anything! We can't belief it.

That's all.

thx mike
//Add this to "OnDomLoad" event