What is OpenDeploy doing.

All:

I have a multi legged OpenDeploy deployment.

Some leg logfiles sit for several minutes with its tail as follows (or similar):

[PHP]Directories deployed : 0 Files deployed : 165 Links deployed : 0
Directories failed : 0 Files failed : 0 Links failed : 0
Directories deleted : 0 Files deleted : 0 Links deleted : 0

LIB: 2010-04-16 05:13:21 File Content transferred: 111882712 bytes
ENG: 2010-04-16 05:13:21 EDT GMT-04:00 Thread-18447 After CLEANUP, res=1
ENG: 2010-04-16 05:13:21 EDT GMT-04:00 Thread-18447 exiting cleanup
ENG: 2010-04-16 05:13:21 EDT GMT-04:00 Thread-18447 entering task-done
ENG: 2010-04-16 05:13:21 EDT GMT-04:00 Begin getInfo request if required
[/PHP]

Other legs look like this:

[PHP]ENG: 2010-04-16 05:24:57 EDT GMT-04:00 Thread-18492 After PRECOMMIT, res=0
ENG: 2010-04-16 05:24:57 EDT GMT-04:00 Thread-18492 exiting dotask
ENG: 2010-04-16 05:24:57 EDT GMT-04:00 Thread-18492 entering awaiting-commit
ENG: 2010-04-16 05:24:57 EDT GMT-04:00 Thread-18492 exiting awaiting-commit
ENG: 2010-04-16 05:24:57 EDT GMT-04:00 Thread-18492 entering commit
LIB: 2010-04-16 05:24:57 *** Proceed
LIB: 2010-04-16 05:24:57 *** fFailedCommitFromEngine 0[/PHP]

The overall job log file syas:

[PHP]ENG: 2010-04-16 05:24:56 EDT GMT-04:00 Waiting for 12 children to complete phases[/PHP]

It would appear that all files have been transferred but OpenDeploy hangs for several minutes trying to complete the deployment.

is anyone aware of any common causes for this situation? What is OpenDeploy doing? Is there any way to speed up this process.

I am working with OpenDeploy 6.2.1 on Solaris.

Many Thanks

Find more posts tagged with

Comments

Migrateduser

Is your deployment transactional?

dazzlad

Yes - it is transactional.

After further investigation we figured out that we had two deployments of the same kind running which seems to confuse the **** out of OD.

OD is configured to reject new deployments if one of the same type is running but as we were passing an instance parameter OpenDeploy took them to be different deploys.

For now we have stopped more than one instance of our workflow running at once pending a better solution.

Thanks.

Adam Stoller

Yes - it is transactional.

After further investigation we figured out that we had two deployments of the same kind running which seems to confuse the **** out of OD.

OD is configured to reject new deployments if one of the same type is running but as we were passing an instance parameter OpenDeploy took them to be different deploys.

For now we have stopped more than one instance of our workflow running at once pending a better solution.

Thanks.

You want to look at information regarding serializeDeploymentsBy* and pathRegistryChecking in the manuals, and also try to access KB 61497 on the Support site.

Essentially the above is for enabling receiver-side queuing of otherwise-concurrent transactional deployments to the same target area.