reliability of GUI replication status

[nojohnny101]

I was wondering if someone could speak to the reliability of the replication status that the GUI reports. I have a periodic snapshot setup to occur once daily and then for replication to a backup-freenas box to occur at night once daily.

I have read on the forums here that one should check the validity of the replication being performed by going to the backup box, clone a recent snapshot, mounting it, and then checking to make sure the data is there that has recently changed.

Question is this: if the GUI reports "up to date" on the replication status and no other errors are occurring on either boxes, can i forgo having to manually check to make it is actually replicating? I have come to the impression that a properly setup freenas box with email reporting and such should not need constant or even regular monitoring, that one should trust the system to let the admin know when there is a problem. Veterans on here have said they sometime don't login to their freenas GUI for months or even a year at a time unless updating is needed.

I would like thoughts on this or if others have had experience where replication said everything was in good working order but then they came to find out by checking manually that it the box wasn't actually replicating successfully.

Thanks for your thoughts!

 

[depasseg]

I rely on the "up to date" message and the lack of a "replication failed" email. As a data point, if my backup box is down (for an update for instance), I tend to get replication failed messages, so I have confidence that lack of message is goodness.

 

[Stux]

You would want to check at least once.

You wouldn't be the first person who setup a non-recursive snap shot and then successfully replicated it.

Or a recursive snapshot with non recursive replication.

In the above cases only the first level directory would contain data.

 

[snaptec]

Or even no data in weird situations.
If the first replication works, everything After normal also.


Gesendet von iPhone mit Tapatalk

 

[nojohnny101]

Thank so for the replies and thoughts everyone.

I have a girl understanding of recursive and non-recursive so I shouldn't run into that mistake. Checking manually one time makes sense and then relying on replication failed emails after that seems sensible.

@depasseg i have also received "replication failed" emails before after restoring a confit file and the SSH key was missing for some reason on the pull box. So I do feel confident enough that when there is a failure, it will be reported.

Thanks everyone. If anyone else sees this thread and wants to chime in feel free.

 

[depasseg]

You wouldn't be the first person who setup a non-recursive snap shot and then successfully replicated it.

It's already been done.
https://forums.freenas.org/index.ph...nsomware-zfs-rollback-having-no-effect.43045/

 

[nojohnny101]

@depasseg yea i saw that the other day and that is what got me thinking!