Running /usr/openv/db/bin/nbdbms_start_server seems to have brought it
back online, there's no information obviously missing, and the various
tools are content to speak to the newly spawned NB_dbsrv, so far as I
can tell.
This seems like a memory leak edge case kind of thing (we weren't doing
anything particular exciting to the EMM database today).
Fwiw, this is under HP-UX 11iv2 (11.23) on a PA-RISC system.
Sigh...
--
gabriel rosenkoetter
Radian Group Inc, Unix/Linux/VMware Sysadmin / Backup & Recovery
gabriel.rosenkoetter AT radian DOT biz, 215 231 1556
-----Original Message-----
From: Rosenkoetter, Gabriel
Sent: Wednesday, April 16, 2008 2:34 PM
To: veritas-bu AT mailman.eng.auburn DOT edu
Subject: Anyway had NB_dbsrv just up and die on 6.5.1?
vmoprcmd -d yields "database server is down", similar responses for
anything else that would want to hit EMM. Google pointed up
http://seer.entsupport.symantec.com/docs/287092.htm, but I'm not having
exactly that same problem, because NB_dbsrv actualy is NOT running, and
today's vault-managed hot catalog backup failed when it tried to connect
too.
Is it theoretically safe just to start NB_dbsrv back up again, like I
would have before with bprd and bpsched? Looking at a system that still
has a running instance, it looks like there're supposed to be a whole
mess of command line arguments...
--
gabriel rosenkoetter
Radian Group Inc, Unix/Linux/VMware Sysadmin / Backup & Recovery
gabriel.rosenkoetter AT radian DOT biz, 215 231 1556
_______________________________________________
Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
|