Unsolved

This post is more than 5 years old

32 Posts

1743

December 7th, 2010 03:00

Domino backup issue (Unable to transfer directory index to xdr stream)

Hello,

We try to backup AIX domino servers with networker (migration on track...).

The last backup (the last try should I say), failed after 2TB backuped.

Messages are:

savegrp: suppressed 2750 lines of output.
Unable to transfer directory index to xdr stream
The operation was aborted
The operation was aborted
NOTES:_3: Failed to wait on the notes save synchronization mutex with id 27263018.(22)
NOTES:_3: Call to release lock on the synchronization mutex with id 27263018 resulted in failure.(22)
The operation was aborted
NOTES:: Failed to wait on the notes save synchronization mutex with id 27263018.(22)
NOTES:: Call to release lock on the synchronization mutex with id 27263018 resulted in failure.(22)
The operation was aborted
NOTES:_1: Failed to wait on the notes save synchronization mutex with id 27263018.(22)
NOTES:_1: Call to release lock on the synchronization mutex with id 27263018 resulted in failure.(22)

I think we should investigate with 'Unable to transfer directory index to xdr stream', but at this time, nothing found.

Thanks

Christophe

116 Posts

December 7th, 2010 03:00

Hi Christopher

Verify the system clocks for the NetWorker server and the NetWorker client must be desynchronized.

Verify compatibility guide too.
Regards,

Luis Rogerio

32 Posts

December 7th, 2010 04:00

Thank you, but servers are synchronized, and compatibility is ok.

I will launch a new backup this afternoon with higher debug mode (3 ?)

Any other idea is welcome

4 Operator

 • 

14.3K Posts

December 7th, 2010 16:00

I think the error itself seems to suggest module is having some trouble to process file list on client towards the server.  I believe both mutex and xdr messages seem to indicate the same thing, but they do not give any more clue.

If this is something that happens all over again, and there is no additional information in Domino and backup application log, debug 3 should be good start.  Are you using NML or NMDA?

32 Posts

December 7th, 2010 23:00

Hi,

This is an extract from a log:

(pid = 8323140) (date = 07.12.10 21:01:23) createObj: Creating object NOTES:/data/Mail/RES01/

(pid = 8323140) (date = 07.12.10 21:01:38) System error: XBSA-1.0.1 LNMs_2010.Build.231 8323140 Tue Dec  7 21:01:38 2010  _nwbsa_close_create_obje

ct_session: System detected error due to: savefile_fini() failure, typically caused by bad network hardware (NICs, routers, etc). Operation aborte

d

(pid = 8323140) (date = 07.12.10 21:01:38) Unable to transfer directory index to xdr stream

and this in an other:

(pid = 6881526) (date = 08.12.10 00:05:32) createObj: Creating object NOTES:/
(pid = 6881526) (date = 08.12.10 00:05:32) createObj: Creating object NOTES:/data/
(pid = 6881526) (date = 08.12.10 00:05:32) createObj: Creating object NOTES:/data/Mail/ABJ01/
(pid = 6881526) (date = 08.12.10 00:05:32) createObj: Creating object NOTES:/data/Mail/ABJ02/
(pid = 6881526) (date = 08.12.10 00:05:32) createObj: Creating object NOTES:/data/notesdata/ARCHIV_CSLD/
(pid = 6881526) (date = 08.12.10 00:05:32) createObj: Creating object NOTES:/data/notesdata/GroupCalendar/
(pid = 6881526) (date = 08.12.10 00:05:32) Backup of connecting directory (browselist) failed

and this (on a lot of databases)

(pid = 8061046) (date = 07.12.10 21:54:41) Error returned for /data/Mail/Mail01/XXXXXX.nsf by Notes API, error = 525 (This database is currently
in use by another person or process, and cannot be accessed at this time.  In order to share a Notes database, it must be accessed via a Domino S
erver by all users of the database.)
(pid = 8061046) (date = 07.12.10 21:54:41) beginSave: Skipping database /data/Mail/Mail01/XXXXXXX since NSR_SKIPDBERRORS is set.
(pid = 8061046) (date = 07.12.10 21:54:41) beginSave failed(211).

should we really use the NSR_SKIPDBERRORS option ??

4 Operator

 • 

14.3K Posts

December 8th, 2010 01:00

If DB is in use you can't back it up so this is why skip is in use.

Back to original problem, it seems as if communication between client and server has been broken during index entry creation (at least I'm reading it that way).  If this is an error which repeats itself try to see if:

- this fails after exactly the same amount of time since backup started

- if timeout setting on client for TCP session is killing idle session

- if you have fw between check fw log for any drops during backup session

32 Posts

December 21st, 2010 06:00

Hi

I answer to myself.

For the 'database in use' problem, it was caused by the fact that some symbolic links were used by domino to access databases, but domino does not support symbolic links.

For the aborting backup, it seems it has been fixed using the 'inactivity limit' of the group (set to 0, infinite, instead of 30 by default).

Thanks to everyone for the help.

Christophe

No Events found!

Top