Decreasing restorePointLimit and deleting backups in 7.2

Lowering the restorePointLimit our vbr backup config file seems to have no effect on the amount of backups actually retained. It is always the highest of whatever previous value has been set. Is there some other method we should be using to lower the number of restore points kept?

 

vbr.py --task listbackup --showconfig
{ ... 'restorePointLimit': 2, ... }
backup epoch objects hosts(nodes) file_system_type
all_backup_20160516_080021 30921636 ... [Linux]
all_backup_20160515_080021 30849523 ... [Linux]
all_backup_20160514_080021 30776673 ... [Linux]

As you can see, 3 backups, but with a config of 2 restore points.

 

Also, since 7.2 I can't find a way to manually remove backups. The entry Removing Backups from the 7.1 docs is not present in the 7.2 documentation, and the file structure has changed making the 7.1 steps invalid. Is there still a way to delete older backups?

Comments

  • Nothing? Seems harsh to remove documented functionality from 7.1 and not replace it?

     

    I have tried deleting all backups and runnint init again, but it just goes back to 3 restore points. The backup config on all nodes only says 2. Is it stored somewhere else as well?

     

     

    We get constant disk space issues as our backups insist on taking 50% more restore points than needed.

  • Hi Andy

     

    Running "select remove_database_snapshot('<snapshot_name>'); to remove  backup should help.

     

    After performing the same please verify it from database_backups and dc_backups for successfull removal along with listbackup output.

     

    Thanks

    Rahul Choudhary

     

     

     

     

  • (Looked this post up today, as again, I needed to delete some backups - seems strange this is such a difficult thing to get any info on!)

     

    I can't remove snapshots as there are not any to remove? Should there be? My understanding of snapshots is that they are held just during the backup process so a consistent fileset can be rsync'd off to the backup server, and then they are removed.

     

    It's the backups on the backup server I need to remove. Since 7.2, Vertica no longer holds the data in individual folders so they can't just be deleted.

     

    Of course, this wouldn't matter so much if Vertica respected the restorePointLimit after being decreased, but I would settle for a solution to either one of these issues for now.

  • restorePointLimit specifies the number of historical backups to retain in addition to the most recent backup. If you set restorePointLimit=2, there will be 2 historical backups, in addition to the most recent backup, for a total of 3 backups.

     

    Standard archive removal method is not available in 7.2.

     

    What version of the Vertica you are running on ?

     

    7.2.3 (SP3) has a feature to remove snapshots.

     

    $ vbr -t remove -c configfile.ini --archive timestamp

  • Thanks for the reply.

     

    We're on 7.2.2 at the moment so I guess we are stuck until we upgrade.

     

    Yeah, we eventually figured out restorePoints was n+1. I have a feeling that in 7.1 though this wasn't the case? The docs for 7.2 also seem refer to the older behaviour. Specifically: 

     

    "For instance, if the restorePointLimit = 5, and five archives exist, running the vbr --task backup utility again deletes the backup with the oldest date_timestamp, when vbr completes the current backup command."

     

    Either way, we managed to progress with our issue. Thanks.

Leave a Comment

BoldItalicStrikethroughOrdered listUnordered list
Emoji
Image
Align leftAlign centerAlign rightToggle HTML viewToggle full pageToggle lights
Drop image/file