Odd Replication Behavior

[nojohnny101]

I am having not really a problem, but just something I don't understand.

I recently moved my backup FreeNAS server (see sig) to a remote location and have everything setup and working fine. I can confirm SSH is working.

On the first night it was to backup, I got an error similar to this:

Hello,
The replication failed for the local ZFS tank/family while attempting to
apply incremental send of snapshot auto-20170208.0100-6m -> auto-20170209.0100-6m to domain.duckdns.org

I thought maybe it was just a dropped internet connection during the night either on my end or the remote end so I ignored it.

2nd night, no error messages.

3rd night (last night). same error message as above.
To confirm, I checked the data on the remote and everything is there that had changed, accessible and browsable.

So my question is, has anyone gotten these errors messages despite replication completely successfully. Strangely enough, the GUI has always had "up to date" on the local side as the status even when it said it failed those two nights.

Any thoughts? Anyone seen this behavior?

 

[PhilipS]

I get these whenever (I assume) there is a small network glitch on my VPN/WAN during replication. The replication script must retry pretty quick, as by the time I check it, everything is fine.

 

[nojohnny101]

After your comment I went back and looked at the network activity around the time of replication and noticed that it did drop off after only 20 minutes. It then waited for 1 hour after that and the network activity picked up to replication levels and continued for exactly 1 hour.

I think you're correct in that some type of WAN interruption happened. Interestingly enough, the time stamp of the "replication failed" email was the exact minute the second attempt started (where it ran for 1 hour).

Thanks for your thoughts. If anyone has more knowledge about how replication handles network interruptions and can confirm this seems to fit that description, it would be much appreciated!