StartupChecksJob Stuck on start-bitbucket.sh

Question

Using BitBucket Server 6.4.2 on an EC2 instance after a vanilla installation. No repositories created yet. After stopping BitBucket with stop-bitbucket.sh, restarting using the startup script appears to be stuck on:

INFO [Caesium-1-2] c.a.b.i.s.i.jobs.StartupChecksJob Running startup jobs for search

Have attempted multiple times and this job does not fail or timeout. Running the startup with --no-search causes this job to move into a cycle of "fail > reschedule." Original goal was to setup HTTPS, but this issue has prevented any progress on that front.

SHORE Dev · Answer

For future reference: I'm not exactly sure what the solution to my problem was.  Some steps I performed were:

* I deleted the search index
* I left BitBucket running overnight
* I rebooted the reverse proxy

I'm not sure which one of these things solved the problem.  I half suspect that there was something strange happening with the reverse proxy and I can't recall if I'd actually tried to access the BitBucket machine directly (ie. NOT through the reverse proxy).

In any case - my problem seems to be solved.  The Atlassian support were very helpful.  They also suggested that I could enable debugging logging by setting the following lines in the $BITBUCKET_HOME/shared/bitbucket.properties file and then restart Bitbucket:

logging.logger.ROOT=DEBUG
logging.logger.com.atlassian.bitbucket=DEBUG

I didn't actually need to do that in the end, but it should come in handy debugging issues in the future.

SHORE Dev · Answer

I've found the same problem.  We're running BitBucket Server v6.10.1 on Docker.  I recently did a `apt-get upgrade' on Ubuntu 18.04.1 and upgraded docker to v19.03.12 and I get stuck at the same place:

INFO [Caesium-1-1] c.a.b.i.s.i.jobs.StartupChecksJob Running startup jobs for search

I disabled elastic search and found it was then stuck in the "fail -> reschedule" cycle you describe...

Did you find out what the solution to the problem was?