unable to replicate datasets

[freenas1]

hello all,

i have two Freenas 9.2.1.7.
On the PUSH host, i have the following datasets:
newpool/ACCOUNTS
newpool/ACCOUNTS/user1
newpool/ACCOUNTS/user2
...
newpool/ACCOUNTS/user500

on the PULL host, the dataset is pool/backups

In Periodic Snapshot Tasks, i have a task for newpool/ACCOUNTS (which recursive checked), and i have snapshots:
# zfs list -r -t snapshot newpool/ACCOUNTS/user1
NAME USED AVAIL REFER MOUNTPOINT
...
newpool/ACCOUNTS/user1@auto-20141112.1200-2w 14.8M - 68.7G -
newpool/ACCOUNTS/user1@auto-20141113.1200-2w 9.33M - 68.7G -

but PUSH host displays the error:
CRITICAL: Replication newpool/ACCOUNTS -> pull.domain.org failed: cannot receive: specified fs (pool/backups) does not exist

and the dataset exists:
# zfs list pool/backups
NAME USED AVAIL REFER MOUNTPOINT
pool/backups 144K 10.7T 144K /mnt/pool/backups

I did the terminal test:
# zfs send newpool/ACCOUNTS@auto-20141031.1200-2w | ssh -i /data/ssh/replication pull.domain.org zfs receive pool/backups/ACCOUNTS@auto-20141031.1200-2w

and it's ok:
pull.domain.org# zfs list -r pool/olympe
NAME USED AVAIL REFER MOUNTPOINT
pool/backups 2.51M 10.7T 152K /mnt/pool/backups
pool/backups/ACCOUNTS 2.36M 10.7T 2.36M /mnt/pool/backups/ACCOUNTS


Anybody can help? Is there a log file about this zfs replication more detailed than that?

thanks in advance,

 

[freenas1]

i don't understand, replication happened one time, 2 days ago, and never rehappen since, is it normal?
Is there a log file where i can find a log about this zfs replication?
i'm interested about time and speed of the replication.

thanks

 

[cyberjock]

i don't understand, replication happened one time, 2 days ago, and never rehappen since, is it normal?
Is there a log file where i can find a log about this zfs replication?
i'm interested about time and speed of the replication.

You tell us. If you scheduled replication for once a month, that might be normal. If you scheduled it for hourly 7 days a week, something is wrong.

 

[freenas1]

You tell us. If you scheduled replication for once a month, that might be normal. If you scheduled it for hourly 7 days a week, something is wrong.

On the PUSH host, in the "Add replication" window, i don't see something about schedule, only "Begin" and "End".
In the "Periodic Snapshot Tasks", the dataset "newpool/ACCOUNTS" has the Frequency "every 1 hour". Certainly it's a mistake, and i've change it to -> 1 day, and now it seems ok.
The only thing is that i'm surprised to not receive an email when the replication is successful?
Does the replication reside in /usr/local/www/freenasUI/tools/autorepl.py? because, in /var/log/cron, i see calls to /usr/local/www/freenasUI/tools/autosnap.py, but no trace for autoreply.pl

thanks for your reply, it helped me a lot!

 

[cyberjock]

Well, with that setup (and assuming the timeframe of the snapshots is also within the begin and end times) you'd expect a snapshot to be created every day, as scheduled, and then replication to immediately pickup. If the snapshot time and replication begin/end times don't coincide then the autorepl.pl should perform a replication at some point between the begin and end (it is not required to run at the begin time).

If you look at the replication settings page it will show the status of the replication in 9.3 (I forget if later 9.2.1.x releases have this or not).

I will say that if you manually transferred a snapshot you'll probably have to check the box that says something like "initialize remote side once" as the replication manager doesn't like it when you manually replicate snapshots. Just letting you know that if you are experimenting you can create new problems in that way. ;)