ralus agent on solaris sun cluster

[bturnbow]

We have a solaris 10 sun cluster and have installed the ralus agent.
The conenction gets authenticated between backup exec server 12.5 and the agent but when we try to schedule a job all we see is the root directory.
Running logging console we get teh following messages

6 Mon May 18 09:52:39 2009 : FS_BlowOutMachine: processing as local machine
6 Mon May 18 09:52:39 2009 : FS_BlowOutMachine: returning 0
6 Mon May 18 09:52:39 2009 : ====> VX_ChangeDir
6 Mon May 18 09:52:39 2009 : ====> VXX_ChangeDir : apath : , asize : 1
6 Mon May 18 09:52:39 2009 : ====>ChangeDir : psize : 1
6 Mon May 18 09:52:39 2009 : vx_stat_by_name: Getting stat of /
6 Mon May 18 09:52:39 2009 : ===> gfsp vxms plugin: memory allocation failed
6 Mon May 18 09:52:39 2009 : ===> gfsp vxms plugin: memory allocation failed
6 Mon May 18 09:52:39 2009 : ===> vf_newpgn: couldn't initialize libgfsp.so
6 Mon May 18 09:52:39 2009 : ===> ufs vxms plugin: memory allocation failed
6 Mon May 18 09:52:39 2009 : ===> ufs vxms plugin: memory allocation failed
6 Mon May 18 09:52:39 2009 : ===> vf_newpgn: couldn't initialize libufs.so
6 Mon May 18 09:52:39 2009 : Function vfmStatFileName, VxMS resolved the error code to this message: object not claimed by any plugin
6 Mon May 18 09:52:39 2009 : vx_stat: objname is / returned 2059
6 Mon May 18 09:52:39 2009 : vx_open_by_name: Opening object name : /, openflags : B
6 Mon May 18 09:52:39 2009 : ===> gfsp vxms plugin: memory allocation failed
6 Mon May 18 09:52:39 2009 : ===> gfsp vxms plugin: memory allocation failed
6 Mon May 18 09:52:39 2009 : ===> vf_newpgn: couldn't initialize libgfsp.so
6 Mon May 18 09:52:39 2009 : ===> ufs vxms plugin: memory allocation failed
6 Mon May 18 09:52:39 2009 : ===> ufs vxms plugin: memory allocation failed
6 Mon May 18 09:52:39 2009 : ===> vf_newpgn: couldn't initialize libufs.so
6 Mon May 18 09:52:39 2009 : Function vfmOpenFileName, VxMS resolved the error code to this message: object not claimed by any plugin
6 Mon May 18 09:52:39 2009 : vx_open_by_name: Error : 2059
6 Mon May 18 09:52:39 2009 : vx_open_by_name: new handle =0
6 Mon May 18 09:52:39 2009 : VX_FindFirstFile: vx_open_by_name failed for path '/' errno=-536836984
6 Mon May 18 09:52:39 2009 : DSEnumerator failed with -536836945:
1 Mon May 18 09:53:07 2009 : Stopping remote agent due to receipt of 'SIGTERM'

We have been unable to finsd anything related to this message.
any ideas?

[GaganSingh]

Hello,
Not sure if you have resolved this or not but I ran into similar situation on the solaris 10 server.  All I was getting was the root but nothing under it.  It turned out to be a bad entry in the /etc/mnttab file which should have been cleared upon reboot.

After clearing the entry, I was able to start the client with no issues.

Hope it helps.

Thanks

Regards,

Gagan

[bturnbow]

Thanks for the sugestion.
I do not see anything strange in the file , here it is.
what excactly di you have ?

thanks

Brian

root@pecbe1 / # cat  /etc/mnttab
/dev/md/dsk/d0  /       ufs     rw,intr,largefiles,logging,xattr,onerror=panic,dev=1540000      1253693621
/devices        /devices        devfs   dev=54c0000     1253693587
ctfs    /system/contract        ctfs    dev=5500001     1253693587
proc    /proc   proc    dev=5540000     1253693587
mnttab  /etc/mnttab     mntfs   dev=5580001     1253693587
swap    /etc/svc/volatile       tmpfs   xattr,dev=55c0001       1253693587
objfs   /system/object  objfs   dev=5600001     1253693587
sharefs /etc/dfs/sharetab       sharefs dev=5640001     1253693587
/dev/md/dsk/d60 /usr    ufs     rw,intr,largefiles,logging,xattr,onerror=panic,dev=154003c      1253693621
fd      /dev/fd fd      rw,dev=5780001  1253693621
/dev/md/dsk/d50 /var    ufs     rw,intr,largefiles,logging,xattr,onerror=panic,dev=1540032      1253693628
swap    /tmp    tmpfs   xattr,dev=55c0002       1253693628
swap    /var/run        tmpfs   xattr,dev=55c0003       1253693628
pecbe1:vold(pid590)     /vol    nfs     ignore,noquota,dev=5800001      1253693685
/dev/md/dsk/d90 /global/.devices/node@2 ufs     rw,intr,largefiles,logging,noquota,global,xattr,nodfratime,onerror=panic,dev=154005a    1253693697
/dev/md/dsk/d40 /global/.devices/node@1 ufs     rw,intr,largefiles,logging,noquota,global,xattr,nodfratime,onerror=panic,dev=1540028    1253693716

[Hemant_Jain]

It looks like an issue mentioned in following document:
http://support.veritas.com/docs/325651
Please use the steps mentioned in this document to confirm, it is the same issue. Currently, i don't see any hotfix for this issue, if it is exactly what you are experiencing. So, currently, you may have to use the workaround as mentioned in the document.

Thanks

[Symanticus]

Ok, so I wonder if this BE 12.5 SP2 RALUS agent can actually performs file level restore on Linux or Solaris ?

[sridhar_s]

The root cause of this issue is under cluster environment /etc/mnttab contains an enry say /dev/md/dsk/d90 which is actually not present on the disk.
Nither raw device or character device of corresponding entry present on the system.

If 2 node cluster is configured and /dev/md/dsk/d90 and /dev/md/dsk/d140 are two globar device paths shared, one path is present on a system which is valied is not present on other system.

While loading providers, caching of file sytem informaton from /etc/mnttab takes place. During process providers validates the device paths usign stat() which is failing on this kind of paths. Due to the failure, providers gfsp and ufs doesn't get loaded and on files gets claimed or files are not browsable from BE.

If the machines are rebooted under non-cluster mode the enties from /etc/mnttab contains valid entry which can be processed by the providers. Hence in non-cluster mode, browing and backup will be successful.

Thanks & Regards,
sridhar.