Environment: Solaris 2.6 Master/Media server, NetBackup 3.4 @ patch
J0850482, STK9710 with 10 DLT7K drives.
Every morning I have bduplicate do duplications of Oracle backups from
the previous day starting at 0600 and they typically end around 0700.
More often than not, when bpduplicate is finished, it tries to output
the bpduplicate log to where I specify with -L, but it gets an error 58
when doing this. BTW: I don't have this problem with non-Oracle
duplicates. Here's what I got this morning:
05/07/2001 06:58:15 nbserver oraclient successfully wrote backup id
oraclient_0989212521, copy 2, fragment 1, 105632 Kbytes at 2870.825
Kbytes/sec
05/07/2001 06:58:20 nbserver bpduplicate database backup cannot connect
to nbserver
05/07/2001 06:58:20 nbserver bpduplicate failure logging message to
client nbserver in log
/usr/openv/netbackup/logs/duplicate_Oracle/log:
can't connect to client (58)
05/07/2001 06:58:20 nbserver bpduplicate Cannot start read process on
nbserver: can't connect to client (58)
05/07/2001 06:58:20 nbserver bpduplicate Duplicate of backup id
oraclient_0989212556 failed, can't connect to client (58).
05/07/2001 06:58:21 nbserver bpduplicate database backup cannot connect
to nbserver
05/07/2001 06:58:21 nbserver bpduplicate Cannot start write process on
nbserver: can't connect to client (58)
After this mess, jobs just hang forever because they can connect to a
storage device (status 213). Last week I killed all active jobs, then
restarted them, and it worked. Today, when I restart them I get 213's.
If possible, I'd like to avoid restarting NetBackup because I have
duplicate jobs running.
Any ideas?
RCA
--
Ryan C. Anderson | United Defense L.P.
Unix Administrator | 763.572.6684 (desk)
ryan_anderson AT udlp DOT com | 952.235.9936 (pager)
|