Deployment hanging sometimes

fiquebem

We're running OD 5.6 SP1 + patch on an NT server. We're deploying some log files from there (i.e., NT) to OD 6.0.1 on Solaris 9. The target directory is a TS 6.5 workarea. Two or three times already I had to restart OD on Solaris because that Log deployment seemed to be hanged for several minutes. Normally, it completes in a minute or so. After killing that deployment (by restarting OD on the target -- I don't have access to the NT server), the Log deployment works fine the next time.

The target area is /iwmnt/default/main/OD_Logs/WORKAREA/logs/etc. Would it be better to deploy to /.iwmnt instead?

The deployment of the logs is initiated by a DNR script (that runs on the target [NT]). But the DNR is specified in a configuration file on Solaris. Doing a iwodserverstatus shows only 2 deployments -- the original Sending one invoked on Solaris, and the Receiving of the logs from NT. I can't cancel the original on Solaris with iwoddeploycancel (it's too late for that).

Checking the workarea where the logs are supposed to be written, I see that, instead of several, only one file was deployed -- a *.iwtmp file.

Any thoughts besides lack of disk space : )?

Incidentally, the log deployment configuration file (on NT) has an includePattern that specifies which logs files to send, e.g.:

<includePattern regex="src\.fanout_portal\.*" />

Thanks,

S.

jbyork

Try messing with transactional vs. non-transactional deployments.
Try messing with timeout settings.
Try deploying to a target directory outside of TeamSite.
Verify correct directory permissions and NT vs AIX considerations (such as file naming conventions, etc.).

See if you can recreate the same problem using different settings and different target directories (as described above).

Are there any firewalls between the servers? Have you run any network tests between the two servers. If deployment works sometimes and hangs sometimes, my guess is that there is something weird going on in your network... and OD waits until it hits the timeout.

I think the *.iwtmp file is just a backup which OD creates during transactional deployments. When deployment "hangs" and you have only 1 *.iwtmp file, my guess is that the error condition occurred on the very first file... and nothing proceeded after that.

Good luck!