Jump to content

"Lock out volumes during backup"

Recommended Posts

Ok, thanks. Do you think that the next version can backup mail server from Apple or MySQL without need of stopping the service ?

That is not and cannot be possible. A cyrus (and SQL database) needs to be self-consistent during backup, and the internal cyrus caches have to be flushed to the database. Failure to do this is a guaranteed corruption of the cyrus database on restore. A better approach would be to use something like the freeware mailbfr script here:


which can stop the mail server, checkpoint the cyrus database, etc., then restart the mail server. Then use Retrospect to back up the checkpointed cyrus database. This could be done with "trigger scripts" using the older "Retrospect Event Handler" in Retrospect 6.x and prior; you can do a similar thing here by scheduling the mailbfr script prior to your Retrospect backup of your mail server.


The mailbfr checkpointing only takes an instant, and your secondary MX should handle mailbagging and forwarding for that brief outage.



Edited by Guest
Link to comment
Share on other sites

Two points.


(1) the "except files in memory" qualification guarantees a corrupt cyrus (and MySQL) database even if an instantaneous copy of the files on disk were to be made.


(2) There is one way to make an instantaneous checkpoint of an entire volume, and it has been used by server admins for years (including us) to keep a server up 24/7/365 through backups. That way is to do a "RAID 1 mirror split". Add a member to a RAID 1 array, allow the RAID to rebuild, make services (cyrus, MySQL, etc.) quiescent, split the member from the RAID, bring the services back up. Then back up the split RAID 1 member.


The new version of SoftRAID (4.0), due "real soon now", is scriptable so that this can be done as a part of your server backup, and there is a white paper on this issue on the SoftRAID site.




Link to comment
Share on other sites

But some other backup software can freeze a computer and backup all data at a point of time.

No, some database programs allow commands that will cause them to create a checkpoint of their database; in other cases, a script does exactly what mailbfr does - stop the service (which causes the in-memory database records to be flushed to disk), checkpoint (copy) the database, restart the database program. If you are under the impression that there's another way, you have been misled.



Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Create New...