I have a deploy that’s been running for almost half an hour now. And there is no explainable delay.
The log is as follows
2012-09-21 07:11:12 INFO Begin deployment
2012-09-21 07:11:12 DEBUG Deploying package MessageBus.Server 188.8.131.5238 to tentacle http://10.27.98.10:10933/
2012-09-21 07:30:08 INFO Deployment successful. Tentacle output follows:
2012-09-21 07:28:50 DEBUG Begin deployment of package: MessageBus.Server.184.108.40.20638
What was it doing in the 17 minutes between Begin deployment and Begin deployment of package?
It’s hard to say, but Tentacle does queue tasks on the machine so that only one deployment is run at once. Perhaps there were other deployments or tasks being run on the remote machine at the time?
I have a similar problem, although that big delays only happened when there was another deployment running on the same machine.
In my case, when I test the deployment by running Tentacle.exe locally from command line, it executes all the steps quickly. We have several packages, and each package does include pre-deploy, deploy and post-deploy scripts. Every time a new script is being started, Octopus visibly pauses for some time - like 30-60 seconds. (I did not measure this exactly). In the beginning I did not care much, as it eventually did do the work, but it is now killing the impact of my demoes: people simply get bored when they have to wait that long for the result. I have been requested by my manager to find out what is causing these delays, and how to correct the situation.
Can you suggest what could I do to find out why the tentacle is pausing? Can I produce some kind of trace that will reveal what it is doing during these delay times? It’s most likely some environment/configuration issue, but I need to find out what needs to be corrected.
The log files on the Tentacle machine may be a good place to start looking for clues:
The documentation refers to OctopusDeploy version 2.0
We are using version 220.127.116.118, which seems not to use the same log files as version 2.0
Can you point me to the documentation that matches our version?
Tentacle in Octopus 1.x uses the Windows Event Log - running eventvwr.exe may help. The Windows system logs from the same time period might reveal some more info.
Hope this helps,
I am experiencing same issue with this version of Octopus.
The event viewer does not show any issues. The traces from PowerShell event log only show normal powershell engine startup, but the actual script starts executing 30 to 60 seconds later.
Does powershell attempt to access internet (eg. to lookup CRLs published by Microsoft? just guessing…) when run by Octopus?
Interesting observation about PS startup, thanks for sharing that.
I don’t think CRLs get involved at this point, but user profile loading does - is it possible the Tentacle user’s profile is taking an extended time for PS to load? Probably a few other items in this category - thanks again for the data point.
I understand it isn’t always possible, but the logging improvements in Octopus 2 are huge - if you’re able to consider upgrading I’d recommend it.
Hope this helps!