Robot Tape drive mount verification error, while performing standalone backup. Please help

[prakashsmtr]

Hi All,

We have 10 nodes cluster setup over Ethernet with shared system disk $1$DGA100.

Out of which we have made 2 nodes to boot minimum 'boot -FL E,o'  for  system disk backup.

But while perform standalone backup we are getting below mentioned error on one of the 2 nodes node and other node is working fine  without error as attached.

Both the nodes are connected to same Compaq MSL500 Tape robot unit and access same tape drives, Please help.

Error on one node:error continues without further progress

%MOUNT-I-MOUNTED, STABCK mounted on _$2$MGA1: (SABKUP)

%MOUNT-I-MOUNTED, SYAPRD mounted on _$1$DGA100: (SABKUP)

Backing up SYAPRD...

%SYSTEM-I-MOUNTVER, $2$MGA1: (SABKUP PGC) is offline.  Mount verification in progress.

%SYSTEM-I-MOUNTVER, $2$MGA1: (SABKUP PGC) has completed mount verification.

%SYSTEM-I-MOUNTVER, $2$MGA1: (SABKUP PGC) is offline.  Mount verification in progress.

[Bob Blunt]

From my point of view you've already made some very significant points.  Your systems all have access to the same tape drives from the SAN.  Your cluster setup/configuration is not germane to this discussion because you're not clustered when you're booted in this manner.  Although you're not really using "standalone" backup you're most likely booting from the distribution CD/DVD (or a disk copy of one) so you have a lot more flexibility than you would with true standalone backup.  In any case I'd be looking at the following suggestions:

Use the most recent distribution CD/DVD you have available, most certainly V8.3 or newer (it wasn't fully clear if your systems were Alpha or Integrity).

Based on your configuration you should only boot ONE system at a time to perform backup of this type.  None of the other nodes in the cluster should be booted.

There shouldn't be any limitation which nodes you could use...unless you've got the tape library on the SAN zoned so only some systems can see them.

Validate the SAN connection from the hosts to the library and individual drives.  There shouldn't be any reason that one node can see the tapes and library "better" than any others.  Make sure that zoning is setup for the OpenVMS systems.

Try different nodes to see if you get any changes in accessability.  In your setup, with the caveat above, all nodes should have equal access to library, tape or disk.

bob

[prakashsmtr]

Thanks Bob for the response....

Below are my latest findings for this issues.

I found that HBA card which is connected to robot unit on this problem node, reported an error while initializing the system, as below, and I have now fixed Topology error back to Fabric as per Google search finding....

So I hope this is going to fix my problem...My standalone backup is scheduled for next week 15/07/213...So I will update on the status next week..Fingers crossed.

P00 BIACPR> init
Initializing...
Initializing pka ega egb egc dqa dqb pga pgb open fibre pgb0.0.0.2.2
pgc pgc0.0.0.2.3 - Nvram read failed.
....
P00 BIACPR> wwidmgr -show adaptor
item    adapter                   WWN                  Cur. Topo  Next Topo
[ 0] pga0.0.0.1.1         2000-0000-c92c-ee9d            FABRIC     FABRIC
[ 1] pgb0.0.0.2.2         2000-0000-c92c-d6b4            FABRIC     FABRIC
pgc0.0.0.2.3 - Nvram read failed.
[ 2] pgc0.0.0.2.3         2000-0000-c92c-1b9b            FABRIC     UNAVAIL
[9999] All of the above.
P00 BIACPR> wwidmgr -set adapter -item 2 topo fabric
pgc0.0.0.2.3 - Nvram read failed.
Reformatting nvram
P00 BIACPR> wwidmgr -show adaptor
item    adapter                   WWN                  Cur. Topo  Next Topo
[ 0] pga0.0.0.1.1         2000-0000-c92c-ee9d            FABRIC     FABRIC
[ 1] pgb0.0.0.2.2         2000-0000-c92c-d6b4            FABRIC     FABRIC
[ 2] pgc0.0.0.2.3         2000-0000-c92c-1b9b            FABRIC     FABRIC
[9999] All of the above.
P00 BIACPR> init
Initializing...

[prakashsmtr]

HI All,

I still have no luck with the above NVRAM fix,  Below are latest update on our setup.

1. Verfied the zoning at storage EVA8000 end  all  3 nodes  & tape robot NSRs are in  seperate proper zone as mentioned below BIACPR is the node causing problem . Verified through SDA FC NAME, all paths are valid.

zone:  TOS_TAPES

                BIABPR_TAPE; AXPDA_TAPE; BIAEPR_TAPE; NSR1_port0; NSR2_port0;

                BIACPR_TAPE; BIAAPR_TAPE; BIAEPR_TAPE1

2. This node(its hba-FGC)  is causing problem not only in stand alone backup but also after booted and even in cluster.

 Tested for image backup of one disk  when booted, backup happens without error (unlike stand alone where it thows error as in my  1st post - PGC goes offlive for mount verification) but exterme/dead  slow  like 10 mb file take 15 mins to backup.

Could any body advise why there is extreme slowness in backup to tape on only one node , when all the nodes are of same configure with HBA of same firmware version....

Is therer any command and utilities to find the reason of slowness of backup...Please help.

I also found some 12 CRC error on SAN port for this  node HBA..

my final thought is to swap the the FC ports b/w working and problem one, to test FC port problem.

Thanks in Advance

Prakash

[prakashsmtr]

Problem resolved,

We replaced the FC cable connected from HBA card to SAN switch...which was the culprit.

As the backup rate was  very slow and was throwing the "verification error" at the standalone boot only and not throwing any error  in the full system boot  conversational backup without any light color  status

But showed some CRC error at storage SAN  switch port end and there we got the clue as FC connectivity is having some problem and tried FC cable replace, it worked

%SYSTEM-I-MOUNTVER, $2$MGA0: (SABKUP PGC) is offline.  Mount verification in pro

gress.

 Finally issue is resolved, many thanks for all your help and suggestion.

Thanks

Prakash