Manual backup/restore not working
We are using the AWS cloudformation template for a teamcity server 2019.1.2 with an RDS instance that hosts the DB. All been working well for a year.
Now we want to clone it, basically we did the equivalent of Manual Backup and Restore for the cloudformation stack:
- we backed up the server by creating a new AWS AMI image of the EC2 instance created by the stack
- we snapshotted the RDS instance
- we created a second stack, but edited the cloudformation template so that:
- the RDS would be created from the RDS snapshot,
- the AMI would be the image created earlier instead of the default TeamCity image for the region
Problem: When we login to the new server, every thing looks good, EXCEPT that the Build Queue shows 500 instead of 0 (which it was in step 1), so we don't think we have a valid restore. I read the page Upgrading TeamCity AWS but nothing jumps out except perhaps that we did not log in as superuser first, rather straight out as an LDAP-authenticated user.
Any idea what might be wrong, any files / settings we could look at to see if maybe there is an undocumented step that needs to be done?
Please sign in to leave a comment.
Hi Oliver,
I think this is the first time I ever see this issue, so I cannot really tell straight away what has gone wrong, if something even has. The build queue is stored in the database, and I'm not too familiar with exactly how RDS operates, so I might be a bit confused in here, but most of the process looks correct. Now, given that the build queue is in the database, the relevant point for the previous build queue size should be 2, and not 1. If projects, older build data, etc is everything there, then it seems like the backup should be correct, but I cannot really tell why the 500 entries in the build queue. It would make sense to analyze some of those entries, see when they claim they were added to the queue, and try to track down what they are doing there.
With this in mind, if you cannot find any trace of the issue, I'd first ensure that your backup and RDS snapshots are correct and should have an empty queue, but also that every project, config, user, etc is available, to see whether just clearing the queue would actually bring you back to a working state. If nothing of that helps, I'd recommend getting in touch with us via the "submit a request" button on top of the page and send us some troubleshooting information, such as when were the builds added to the queue, date of backup/restore, etc.