'
[2020-05-16 10:09:19] Deleting old backups...
[2020-05-16 10:09:19] Cleaning stuff up...
[2020-05-16 10:09:19] Removing '.tar' leftovers...
[2020-05-16 10:09:19] Marking backup as finished...
[2020-05-16 10:09:19] Refreshing disk stats...
[2020-05-16 10:09:19] Notifying 'RikoDEV' of the end of the backup...
[2020-05-16 10:09:29] Something went wrong while notifying user.
[2020-05-16 10:09:29] Finished!
```
---
### Author:  [@codinghorror](https://meta.discourse.org/u/codinghorror)
#### Post date: [May 16, 2020, 11:18pm UTC](https://meta.discourse.org/t/corrupted-item-pointer-during-backup-process/151680/2 "2020-05-16T23:18:47Z")
Looks like a corrupt database to me based on the error message?
---
### Author:  [@RikoDEV](https://meta.discourse.org/u/RikoDEV)
#### Post date: [May 17, 2020, 7:46am UTC](https://meta.discourse.org/t/corrupted-item-pointer-during-backup-process/151680/3 "2020-05-17T07:46:53Z")
Is there any way to repair this database without losing data? I have the last backup 2 months ago and I donβt want to lose the created topics.
---
### Author:  [@Falco](https://meta.discourse.org/u/Falco)
#### Post date: [May 17, 2020, 4:25pm UTC](https://meta.discourse.org/t/corrupted-item-pointer-during-backup-process/151680/4 "2020-05-17T16:25:03Z")
> [@RikoDEV](#):
>
> `public.scheduler_stats`
You got lucky that this table contains just sidekiq stats and nothing really important. You can try to remove this table rows and then try to trigger another backup. Hope only this table was affected π€
---
### Author:  [@RikoDEV](https://meta.discourse.org/u/RikoDEV)
#### Post date: [May 17, 2020, 4:45pm UTC](https://meta.discourse.org/t/corrupted-item-pointer-during-backup-process/151680/5 "2020-05-17T16:45:33Z")
The problem is that postgresql crashes when I try to clean this table.
```plaintext
discourse=# DELETE FROM scheduler_stats;
PANIC: corrupted item pointer: 10992
server closed the connection unexpectedly
This probably means the server terminated abnormally
before or while processing the request.
The connection to the server was lost. Attempting reset: Succeeded.
```
---
### Author:  [@Falco](https://meta.discourse.org/u/Falco)
#### Post date: [May 17, 2020, 4:53pm UTC](https://meta.discourse.org/t/corrupted-item-pointer-during-backup-process/151680/6 "2020-05-17T16:53:58Z")
Oh noes. Does a TRUNCATE on this table also fails?
---
### Author:  [@RikoDEV](https://meta.discourse.org/u/RikoDEV)
#### Post date: [May 17, 2020, 4:55pm UTC](https://meta.discourse.org/t/corrupted-item-pointer-during-backup-process/151680/7 "2020-05-17T16:55:59Z")
TRUNCATE works, thanks. π. I will now check if the backup is successful.
EDIT:
Oh no βΉ !
```plaintext
[2020-05-17 16:56:39] pg_dump: Dumping the contents of table "post_search_data" failed: PQgetResult() failed.
[2020-05-17 16:56:39] pg_dump: Error message from server: ERROR: invalid page in block 89 of relation base/16384/22213
[2020-05-17 16:56:39] pg_dump: The command was: COPY public.post_search_data (post_id, search_data, raw_data, locale, version) TO stdout;
[2020-05-17 16:56:39] Unpausing sidekiq...
[2020-05-17 16:56:39] EXCEPTION: pg_dump failed
[2020-05-17 16:56:39] /var/www/discourse/lib/backup_restore/backuper.rb:183:in `dump_public_schema'
```
EDIT 2 π
Fixed using [https://stackoverflow.com/questions/5220344/postgresql-invalid-page-header-in-block](https://stackoverflow.com/questions/5220344/postgresql-invalid-page-header-in-block).
---
### Author:  [@system](https://meta.discourse.org/u/system)
#### Post date: [June 16, 2020, 5:01pm UTC](https://meta.discourse.org/t/corrupted-item-pointer-during-backup-process/151680/8 "2020-06-16T17:01:13Z")
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.