----- Forwarded by Davis Kwong/Menlo Park/IBM on 08/06/2003 05:18 PM -----

Davis Kwong
To: "rkusenet" <rkusenet@sympatico.ca>
08/06/2003 05:19 cc: informix-list@iiug.org, owner-informix-list@iiug.org
PM From: Davis Kwong/Menlo Park/IBM@IBMUS
Subject: Re: onbar hangs(Document link: Davis Kwong)




Hi, I'm not sure which manual is for 9.21, but here's a cut & paste of the
9.4 IBM Informix Backup and Restore Guide which talks about controlling
parallel backup and restore using onbar.

The BAR_MAX_BACKUP parameter specifies the maximum number of
parallel processes that are allowed for each onbar command. Both UNIX and
Windows support parallel backups. Although the database server default
value for BAR_MAX_BACKUP is 4, the onconfig.std value is 0.

To perform a serial backup or restore, set BAR_MAX_BACKUP to 1.
ON-Bar ignores the BAR_MAX_BACKUP parameter for a whole-system
backup because they are always done serially.
onconfig.std value 0
if value not present 4
units onbar processes
range of values 0 = maximum number of processes allowed on system
1 = serial backup or restore
n = specified number of processes spawned
takes effect When onbar starts

To specify parallel backups and restores, set BAR_MAX_BACKUP to a value
other than 1. For example, if you set BAR_MAX_BACKUP to 5 and execute an
ON-Bar command, the maximum number of processes that ON-Bar will
spawn concurrently is 5. Configure BAR_MAX_BACKUP to any number up to
the maximum number of storage devices or the maximum number of streams
available for physical backups and restores.
If you set BAR_MAX_BACKUP to 0, the system creates as many ON-Bar
processes as needed. The number of ON-Bar processes is limited only by the
number of storage spaces or the amount of memory available to the database
server, whichever is less.
The amount of memory available is based on SHMTOTAL. ON-Bar performs
the following calculation where N is the maximum number of ON-Bar
processes that are allowed:
N = SHMTOTAL / (# transport buffers * size of transport buffers / 1024)
If SHMTOTAL is 0, BAR_MAX_BACKUP is reset to 1. If N is greater than
BAR_MAX_BACKUP, ON-Bar uses the BAR_MAX_BACKUP value. Otherwise,
ON-Bar starts N backup or restore processes.

thanks, Davis.




"rkusenet"
<rkusenet@sympati To: informix-list@iiug.org
co.ca> cc:
Sent by: Subject: Re: onbar hangs
owner-informix-li
st@iiug.org


08/06/2003 02:03
PM
Please respond to
"rkusenet"





"rkusenet" <rkusenet@sympatico.ca> wrote in message
news:bgro1s$rbvfo$1@ID-75254.news.uni-berlin.de...
> IDS 9.21.UC4
> Solaris 2.6
>
> EDS takes our backup using onbar. What we have observed is that,
> quite frequently onbar hangs. By hangs it means, onbar processes
> keep running indefinitely. Since the backup is taken daily, all
> subsequent backups also hang until the earlier on is killed.
>
> I have a script to look for onbar process 15 hours after it
> starts, and if it finds one, it sends a mail as follows:-
>
> =========================================
> onbar is still running on flx10
> root 15709 15703 0 00:50:14 ? 0:00 /bin/sh
/opt/informix/bin/onbar -b -L 0
> root 15878 15711 0 00:51:14 ? 0:00 /opt/informix/bin/onbar_d
-b -L 0
> root 15879 15711 0 00:51:14 ? 0:00 /opt/informix/bin/onbar_d
-b -L 0
> root 15873 15711 0 00:51:14 ? 0:00 /opt/informix/bin/onbar_d
-b -L 0
> root 15872 15711 0 00:51:14 ? 0:00 /opt/informix/bin/onbar_d
-b -L 0
> root 15871 15711 0 00:51:14 ? 0:00 /opt/informix/bin/onbar_d
-b -L 0
> root 15874 15711 0 00:51:14 ? 0:00 /opt/informix/bin/onbar_d
-b -L 0
> root 15711 15709 0 00:50:14 ? 0:01 /opt/informix/bin/onbar_d
-b -L 0
> =========================================
>
> The ssyadmin then kills the processes indicated above.
>
> Once they kill the client process, we get this message in online.log
> =======================================
> 15:04:10 Archive on logs02 ABORTED.
> 15:04:10 Aborted by client.
> 15:04:10 Archive on logs03 ABORTED.
> 15:04:10 Aborted by client.
> 15:04:10 Archive on logs04 ABORTED.
> 15:04:10 Aborted by client.
> 15:04:10 Archive on physdbs ABORTED.
> 15:04:10 Aborted by client.
> 15:04:11 Archive on fdbs04 ABORTED.
> 15:04:11 Aborted by client.
> 15:04:11 Archive on gdbs01 ABORTED.
> 15:04:11 Aborted by client.
> =======================================
> In other words, the backup is useless.
>
> Earlier on, it use to be once/twice a month. Now it is once/twice a week.
> What could be wrong with it. Why does onbar hang for 15 hours without
doing
> anything.

Which manual describes onbar commands for 9.21?

It seems it has something to do with parallelism.

current request exceeds parallelism.

solaris log file shows this error:-

> XBSA-1.0.1 6.0.Build.153 15872 Wed Aug 6 00:52:17 2003
> _nwbsa_is_retryable_error: received a retryable network error
> (Severity 0 Number -13): current request exceeds parallelism
> XBSA-1.0.1 6.0.Build.153 15873 Wed Aug 6 00:52:17 2003
> _nwbsa_is_retryable_error: received a retryable network error
> (Severity 0 Number -13): current request exceeds parallelism
> XBSA-1.0.1 6.0.Build.153 15871 Wed Aug 6 00:52:17 2003
> _nwbsa_is_retryable_error: received a retryable network error
> (Severity 0 Number -13): current request exceeds parallelism
> XBSA-1.0.1 6.0.Build.153 15874 Wed Aug 6 00:52:19 2003
> _nwbsa_is_retryable_error: received a retryable network error
> (Severity 0 Number -13): current request exceeds parallelism

I would like to see whether there are options in onbar to turn off
parallelism.




sending to informix-list