View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0000958||bareos-core||[All Projects] General||public||2018-06-01 16:22||2020-02-17 04:30|
|Priority||high||Severity||major||Reproducibility||have not tried|
|Fixed in Version|
|Summary||0000958: Always incremental backups always upgraded to full|
|Description||After upgrading bareos from 16.2 to version 17.2, each "always incremental" backup job is upgraded to a full backup. This happens all the time and the unmodified configuration used to work fine with 16.2.|
Client: Ubuntu 18.04
Server (Director, Storage Daemon): FreeNas 11.1 (Freebsd 11.1-RELEASE-p10)
|Steps To Reproduce||On a fresh installation, create an always incremental job. The job will be upgraded to full during the first backup (expected) and subsequent ones (not expected).|
|Additional Information||I figured maybe the script that upgraded the tables had a problem, so I dropped the whole database and recreated it again and reset bareos to zero (deleting all the stored data and status on every single component including the 2 clients I have). In this way I ruled out problems with the upgrade from one version to another, but the problem remains.|
Looking at the logs I see the confirmation the first full backup upgraded from incremental finished with status OK, but the following same job (which should run as incremental) is upgraded "from Full to Full" and the reason is that a prior failed full job was found in the catalog, but the database table has the right information for the OK status.
I am attaching a sanitized version of the log file with the two Job Ids involved. The rest of the logs belong to the second client failing jobs (that on purpose I turned off) to see what was going on more easily.
|Tags||always incremental, freebsd|
bareos.log (6,342 bytes)
Problem still present in version 19.2.6 and Postgres 11.
The message: "Prior failed job found in catalog. Upgrading to Full." is not very helpful. At least it should show the previous JobId its referencing to and the job status (in my latest test it will show that the JobID is 1 and the status is T).
Is there any way to debug this behavior?