1) Set all the montior switches for my user session
2) Run db2 "get snapshot for all on MYDB"
One of two things will happen next:
1) My db connection will be forced off and the snapshot will not be
returned
2) One of the monitor switches will get turned off ( i.e. the
bufferpool switch) and no data returned for that group.
Has anyone else encountered this or have any idea what might be
causing it? Not sure I can submit this to DB2 support since there is
no error output. I've already scoured this newsgroup for hours
looking for something similar with no luck.
Thanks in advance for your replies.
AntiVirus DBA
A virus perhaps? ;-)
(sorry, couldn't resist)
I suspect DB2 isn't able to perform the full snapshot due to lack of
resources.
Did you try switching on various subsets of the monitors and see what
happens?
Also carefully check the diagnostics and notification logs.
In any case I would open a PMR, "no error output" doesn't imply "no
support".
IBM will probably ask you to run the db2support command, which gathers all
sorts of info not only on DB2 level but also on the OS.
HTH.
--
Jeroen
I'm sure you know that V8 is out of support since early this year.
Do you get anything in db2diag.log? What if you set DIAGLEVEL = 4?
>
> I'm sure you know that V8 is out of support since early this year.
>
Yep, we're in the process of moving off. Trying to get snapshot
statistics in order to plan for the new v9 environment.
--AV DBA
I'm thinking the same thing, but attempts to reproduce in an alternate
environment are unsuccessful. Switches to various subsets seem
successful. But I want it all!
> Also carefully check the diagnostics and notification logs.
>
> In any case I would open a PMR, "no error output" doesn't imply "no
> support".
> IBM will probably ask you to run the db2support command, which gathers all
> sorts of info not only on DB2 level but also on the OS.
That's good to know. Our diaglog has been archived but I will look
there the next time this happens. We did go through a project to
increase the number of spindles the data resides on because of I/O
wait. The system load has been greatly reduced and we haven't seen
the error again since then.