Backup does not work
Hi,
"YouTrack encountered a problem when creating a backup of your database: Blob file /opt/youtrack/data/youtrack/blobs/3/41.blob expected length is 78463 but actual is 86701"
And now I cannot make a backup.
How to check and fix this file?
Please sign in to leave a comment.
Hi!
What YouTrack version do you have?
When I run a backup, an icon is drawn that the backup is being created. But when I refresh the page, the icon is where yesterday's backup was. And in the end, the same backups remain that were before the backup was launched.
Would it be possible to upload your logs folder to our file-sharing service?
/opt/youtrack/logs?
Upload id: 2023_04_28_YT25emgKMfCWW7x1QAaxSUKz (file: yt-logs.tar.gz)
Thank you for the logs!
It looks like a bug. Please subscribe to it to get a notification when it is fixed (it should be fixed during next week). Sorry for the inconvenience!
Ok, thank you!
Hi, same problem here with YouTrack 2023.1.9286, we see this kind of errors in logs:
```
[2023-06-27 11:38:42,289] jetbrains.exodus.ExodusException: Store : 'teamsysstore' data is broken. Can not read blob located at '/data/youtrack/data/youtrack/blobs/6/6d.blob'. Blob property name 'content'. Entity id : '44-79'. Entity type : 'PersistentFile'. Real length 402760 bytes. Expected blob length 17440 bytes.
```
We upgraded to 2023.1.x on 2023-06-01 and those errors started showing up on 2023-06-22.
It seems that some attachment file uploads or displays are failing due to this, as well as backups.
I'm afraid that we loose some data right now...
We can we do?
The solution provided in the issue https://youtrack.jetbrains.com/issue/XD-933/Create-option-to-run-blob-consistency-refactoring-in-case-of-corruption-of-blob-metadata.?_gl=1*1h7ecc*_ga*MTkxNDY2NjA2My4xNjg3NDIxNDky*_ga_9J976DJZ68*MTY4Nzg2NjUxMy4yLjEuMTY4Nzg2NzQ0NS41OS4wLjA.&_ga=2.178484387.647353718.1687866514-1914666063.1687421492 seems to work: we added the option `-Dexodus.entityStore.refactoring.clearBrokenBlobs=true` as described there, and no more errors after restarting youtrack.
Let's hope it doesn't happen again.
Grégory, yes, this issue is related to the database upgrade that was included in 2023.1. The JVM option you used usually helps. We recommend upgrading to the latest available Server version (Get Youtrack), as most of these issues have already been fixed.