Octopus Server Index folder filling up disk

Hi,

We have been Octopus v2.5.8.447 for a couple years without problems. Yesterday it stopped working as the disk was full, so I removed some old backup files and started the Octopus Server service. Before starting the service I had 1.4 GB free, but after starting the service there suddenly was over 10 GB. To be on the safe side, I changed the retention policies to keep fewer old releases.

Today the disk was full again and the serivce stopped. Looking at the disk, I see that 99% of the disk space is used by the Repository folder of Octopus Server, with 56 GB for the Index folder and 40 GB for the Packages folder. I removed some backups to get 1 GB free disk, started the service and now suddenly the disk has 57 GB free. The Index folder now is using only 0.5 GB, why did it grow to 56 GB during the night?

Brgds

Jonas

Hi again,

30 minutes later, the Index folder has grown to 5 GB without anyone doing deploys. Will it consume all of the disk again (55 GB left)?

Brgds

Jonas

Hi,

The Index folder is now less than 300 MB which is good. But now I have a hanging/active “Check Tentacle Health” task that failed due to the full disk, and now I can’t get it to run. I’ve tried to rerun it, cancel it, restart the Octopus Server service but it is still there. It blocks backups and other tasks.

What to do?

Brgds

Jonas

The hanging issue was solved after doing this:

http://docs.octopusdeploy.com/display/OD2/Manually+fail+a+running+or+cancelling+task

Still wondering why Index folder was going crazy though.

Brgds

Jonas

Hi Jonas,

Thanks for getting in touch! I am glad you were able to get everything up and running again with the issues that you have had.
Why the index folder went crazy, I cannot tell you. I have never had that reported before, but it easily could have been a corrupted index.
I wonder do you have very large meta-data in your packages such as release notes? Did you notice one index file was larger than any other?
When you delete from the package directory you need to restart the Octopus service to correctly remake the indexes.

Let me know if you remember anything specifically otherwise we may just have to keep an eye on the size then we can go from there.

Vanessa