Bitbucket Server doesn't start after upgrade to 5.3

Todd Barber August 17, 2017

The installation appeared to go smoothly with no errors reported. However, after starting, Bitbucket is listed as running in the Services application, but the atlassian-bitbucket.log and just quits.

Has a log entry 

2017-08-17 11:08:58,875 INFO [I/O dispatcher 1] c.a.b.s.s.t.DefaultElasticsearchConnectionTester Connected to Elasticsearch version 2.3.1

but the next entry is

2017-08-17 11:08:59,219 INFO [Caesium-1-3] c.a.b.s.j.c.i.HealthCheckRunner Health checks will run again after PT5M

and is just repeating these but the minute check just conitnues to double. So it goes to PT10M, then PT20M

Thanks for any input.

1 answer

0 votes
Dave Chevell
Atlassian Team
Atlassian Team members are employees working across the company in a wide variety of roles.
August 23, 2017

Hey Todd,

 

The logs suggest Bitbucket is running. Did you upgrade from Bitbucket 4.x? If so, you need to migrate your server.xml customisations to bitbucket.properties. If you're running Bitbucket behind a reverse proxy or on a port other than the default, failing to migrate these settings might explain why you don't appear to be able to connect to Bitbucket. For more information on that, check out this article: https://confluence.atlassian.com/bitbucketserver/migrate-server-xml-customizations-to-bitbucket-properties-897811761.html

 

Cheers,

Dave Chevell  

Todd Barber August 24, 2017

So, you are right, it was running. But the server was consumed doing something else. I put in a ticket with support and gathered some logs and they asked me to get threads. But an hour or so later the server had finished whatever and it was reachable through the web. Don't know what caused it and haven't seen any sluggishness since.

Maximilian Heyne [Communardo] October 4, 2018

Hi Todd, Hi Dave,

we tried to update our Bitbucket today from version 4.11. to version 5.12.2 and ran into the same issue. Unfortunately after about two and a half hours we stopped it and rolled back to the old version. 

I do not know, why the indexing (at least this it what it seams it does in the background) is necessary for the update process itself. In my opinion this could run in the background once it is assured that the elastic search service is running. 

Is there a possible workaround, or a way to analyse what it is doing in the background and to monitor the progress?

Cheers, 

Maximilian Heyne

Suggest an answer

Log in or Sign up to answer
TAGS
AUG Leaders

Atlassian Community Events