Bad file leading to SQLBackup invisible hang

RichardB

Slightly odd scenario here, resulting in:

18210 :
BackupMedium::ReportIoError: read failure on backup device 'SQLBACKUP_A7B98D34-4268-4F1C-8BDA-DDD07071EEE3'. Operating system error 995(error not found).

I think what happened was the backup wrote a partial file (possibly disk space issue) which was copied across anyway. What then happens is the job tries to process the file but just hangs, sqbstatus giving nothing, but the agent job still running - and not easily cancellable.

After a bit of mucking about retrying this (with some hair pulling) I eventually got round to running the first file in the sequence alone, resulting in the above message in the logs. By eliminating this file the restore is ok.

Any ideas?

RichardB

Also I get quite a lot of

Error 920: Failed to update backup history table (BACKUPMEDIAFAMILY).
SQL error 1205: Transaction (Process ID 64) was deadlocked on lock resources with another process and has been chosen as the deadlock victim. Rerun the transaction.

When I have simultaneous backups or restores happening.

petey

What is the version of SQBCoreService.exe that is intalled?

RichardB

4.5.0.704

RichardB

While we are here... I also am now getting a problem with teh log backups.

I see that the cluster got bounced while one log was being backed up, then bounced back while a different databases log was being backed up.

A third databases logs are fine, but those 2 just sit doing nothing, and don't back up.

A standard mssql log backup works fine tho. I have restarted the sqlbackup service, to no avail.

RichardB

at this point sqbstatus sits with both the inert processes with 0 in the processed and compressed, while the one that is still functional pops in and out with large numbers or no record (presumably depending on whether its running at the time!)

RichardB

Ah interesting. This also seems to be blocking the differential backup for those dbs too.

petey

An incomplete backup file should not have been copied to the entered folders. Was the copy done via SQL Backup, or another mechanism?

Re error 920, I would suggest upgrading to version 4.6.

To clear the 2 processes that are not doing anything, stop the SQL Backup service for at least a minute before starting it again.