Sick server session
Neil Truby reported an IDS 9.21 Red Hat primary in an Enterprise Replication pair that rejected onmode and onstat commands, then after a reboot refused connections for about 25 minutes while ER initialisation slowly failed. Hari Gupta suggested checking the clock difference between servers and rebuilding replication with cdr delete server. Unresolved.
Auto-generated by Claude from the posts below — may be imperfect; read the full thread.
IDS 9.21 on Red Hat
My client's database session is really not too well. The database server
will not accept memory-type operations, such as onmode -k or onstat -F.
There are also no oninit proceses running. An onstat -g seg shows single
segments, but an ipcs -m shows multiple ones. In most other respects the
server is OK, although it is the primary in an Enterprise Replication pair,
and this proces is affected too, although there are clues that this may be
unrelated.
He's re-booting the Linux server tomorrow so hpefully all will come up
cleanly, but I just wondered if anyone has seen similar behaviour in the
past ....?
thanks
Neil
Having been re-booted, it's still ill:
Now when the server is stopped, it starts up quickly into on-line mode, but
continues to reject connections with the message
18:51:03 (1805) connection rejected - no calls allowed for sqlexec
18:51:03 listener-thread: err = -27002: oserr = 0: errstr = : Noconnections are allowed in
for about 20 minutes until finally ...
19:14:05 CDR queuer initialization complete
19:14:05 DDR Log Snooping - Snooping started in log 217
19:14:06 CDR connection to server lost, id 2, name <g_host2>Reason: attempt to connect failed
(This connect is failing because the replcation server is shut down).
So it seems that the Enterprise replication initialisation is taking
inordinately long - about 25 minutes - to fail.
Any observations gratefully received.
Neil Truby t:01932 724027
Director m:07798 811708
Ardenta Limited e:neil.truby@ardenta.com
"Neil Truby" <neil.truby@ardenta.com> wrote in message
news:bpmf18$1p43ib$1@ID-162943.news.uni-berlin.de...
> IDS 9.21 on Red Hat
>
> My client's database session is really not too well. The database server
> will not accept memory-type operations, such as onmode -k or onstat -F.
> There are also no oninit proceses running. An onstat -g seg shows single
> segments, but an ipcs -m shows multiple ones. In most other respects the
> server is OK, although it is the primary in an Enterprise Replication
pair,
> and this proces is affected too, although there are clues that this may be
> unrelated.
>
> He's re-booting the Linux server tomorrow so hpefully all will come up
> cleanly, but I just wondered if anyone has seen similar behaviour in the
> past ....?
>
> thanks
> Neil
>
>
Neil,
1. I wud like to check time on both server. Make sure time diff. should not
be more than 4-5 min.
2. How big is secondary database ? and can you afford to rebuild replication
? If yes, you need to use cdr delete server on primary and target both
before rebuilding replication.
Hari
"Neil Truby" <neil.truby@ardenta.com> wrote in message
news:bpnsrt$1kiag6$1@ID-162943.news.uni-berlin.de...
> Having been re-booted, it's still ill:
>
> Now when the server is stopped, it starts up quickly into on-line mode,
but
> continues to reject connections with the message
>
> 18:51:03 (1805) connection rejected - no calls allowed for sqlexec
> 18:51:03 listener-thread: err = -27002: oserr = 0: errstr = : No> connections are allowed in
>
> for about 20 minutes until finally ...
>
> 19:14:05 CDR queuer initialization complete
> 19:14:05 DDR Log Snooping - Snooping started in log 217
> 19:14:06 CDR connection to server lost, id 2, name <g_host2>> Reason: attempt to connect failed
>
> (This connect is failing because the replcation server is shut down).
>
> So it seems that the Enterprise replication initialisation is taking
> inordinately long - about 25 minutes - to fail.
>
> Any observations gratefully received.
> Neil Truby t:01932 724027
> Director m:07798 811708
> Ardenta Limited e:neil.truby@ardenta.com
>
> "Neil Truby" <neil.truby@ardenta.com> wrote in message
> news:bpmf18$1p43ib$1@ID-162943.news.uni-berlin.de...
> > IDS 9.21 on Red Hat
> >
> > My client's database session is really not too well. The database
server
> > will not accept memory-type operations, such as onmode -k or onstat -F.
> > There are also no oninit proceses running. An onstat -g seg shows
single
> > segments, but an ipcs -m shows multiple ones. In most other respects
the
> > server is OK, although it is the primary in an Enterprise Replication
> pair,
> > and this proces is affected too, although there are clues that this may
be
> > unrelated.
> >
> > He's re-booting the Linux server tomorrow so hpefully all will come up
> > cleanly, but I just wondered if anyone has seen similar behaviour in the
> > past ....?
> >
> > thanks
> > Neil
> >
> >
>
>
"Hari Gupta" <hariog@yahoo.com> wrote in message news:3fbfd486$1@news.alphalink.com.au... > Neil, > > 1. I wud like to check time on both server. Make sure time diff. should not > be more than 4-5 min. Yes, it is. That's why database 2 has been brought down. They don't seem to have stopped ER though .... Thanks Neil