Hi,
we have a strange HDR-Problem:
OS AIX 5.2 ml 4.0
IDS 9.40.FC5

The system has been working without big problems for a long time.
For synchronisation we have a fast cross -link-connection seperated
from user connections.

Normal Checkpoint duration is some seconds. Sometimes the system is
very busy and then checkpoint duration can be 30 - 60 seconds.

Last weeks we had two situations, where checkpoint duration was more
than 7000 seconds.
after that onstat -g dri showed:
on for secondary
off for primary.

I dont know, if the long CheckPoint is the reason or the result of the
(unknown) problem.

The necessary logs were available on primary but there was no
resynchronisation after that checkpoint and out of sync.
After Stopping and Starting the secondary the systems were synchron
after a while.

The primary has to be high available and we are not able to do some
experiments.

First time when the servers were out of sync, we tried to resync with
an L0-archive on file on primary versus remote device from secondary
(over fast cross link connection , filesize approx. 90 GB) .

After a while we could not start any ps or other process on the
primary (folk failed) and we had wait I/O ~99,99% . WE killed the
restore via remote device. After approx. 30 min the wait I/O was
normal. Then we made an archive on tape, transported the tape to the
secondary, restored and synchronised.

Are there any known problems with remote devices on this OS/IDS
combination.
Are there known bugs in IDS940.FC5 related to HDR ?
I known, that the current 9.40 version is XC9, but there is no chance
for an upgrade without justification in relation to fixed HDR bugs.

Thanks in advance
Rüdiger Papke






is made

On Dec 17, 4:04*pm, ruediger.pa...@t-online.de wrote:
> Hi,
> we have a strange HDR-Problem:
> OS AIX 5.2 ml 4.0
> IDS 9.40.FC5
>
> The system has been working without big problems for a long time.
> For synchronisation we have a fast *cross -link-connection seperated
> from user connections.
>
> Normal Checkpoint duration is some seconds. Sometimes the system is
> very busy and then checkpoint duration can be 30 - 60 seconds.
>
> Last weeks we had two situations, where checkpoint duration was more
> than 7000 seconds.
> after that onstat -g dri showed:
> on *for secondary
> off *for primary.
>
> I dont know, if the long CheckPoint is the reason or the result of the
> (unknown) problem.
>
> The necessary logs were available on primary but there was no
> resynchronisation after that checkpoint and out of sync.
> After Stopping and Starting the secondary the systems were synchron
> after a while.
>
> The primary has to be high available and we are not able to do some
> experiments.
>
> First time when the servers were out of sync, we tried to resync with
> an L0-archive on file on primary versus remote device from secondary
> (over fast cross link connection , filesize approx. 90 GB) .
>
> After a while we could not start any ps or other process on the
> primary (folk failed) and we had wait I/O ~99,99% . WE killed the
> restore via remote device. *After approx. 30 min the wait I/O was
> normal. Then we made an archive on tape, transported the tape to the
> secondary, restored and synchronised.
>
> Are there any known problems with remote devices on this *OS/IDS
> combination.
> Are there known bugs in IDS940.FC5 related to HDR ?
> I known, that the current 9.40 version is XC9, but there is no chance
> for an upgrade without justification in relation to fixed HDR bugs.
>
> Thanks in advance
> Rüdiger Papke
>
> is made

Can you post your ONCONFIG from Primary Server and also Online log
(between the days you had 2 Long Checkpoints) ?

Thanks
prasad