Replication task still in progress and the receiving end has more data the the sending end, huh

[papageorgi]

Hello everyone I am hoping someone can clear this up for me, I seem to be really confused on using the replication task and am thinking of switching over to rsync. I have two servers A and B. I am looking to have server A be what I actually use and B just to be a backup, that's it. Since I have daily snapshots on A I really just want the data on server B and for it to incrementally be sent over to the backup server.

After getting this setup with a few hurdles and having this large amount of data transferring over my 1GBe LAN (with a UniFi 48 port switch) it still is not done a long time later but it does not make sense. The backup server (server B) almost has as much data as the main server (server A) when you look at the jails and iocage, but the actual dataset that I have my data on is larger. I was reading the documentation to see if I am missing something but it seems like I may be missunderstanding the replication settings. I am looking to move the dataset "stor" with it's iocage, jails and "temp" datasets so I selected the "recursively replicate child dataset snapshots," is this the problem?

 

[dlavigne]

Were you able to figure this out?

 

[papageorgi]

Hello dlavigne, not really no, but thank you kindly for asking ... I basically just let it continue after a few restarts (one after a short black out). In the end it took a lot more storage on the backup server and since it seemed like everything was there after looking it over for some time and the original pool is well over 10 years old I decided to erase the old pool. I took a screenshot of the end result when it completed (attached).

I added a few drives to newly recreated pool for more storage and currently I am copying the data back with rsync since I do not want anything but the data itself. Sadly, it is very true about the slower performance. Rsync is running about half the speed that I was getting with replication (zfs send/receive).