JustinClift
Patron
- Joined
- Apr 24, 2016
- Messages
- 287
So, I'm trying to set up replication from scratch between two FreeNAS 11.3-U5 boxes, and it's not going well at all. :(
I keep on getting an error message about `cannot unmount '/var/db/system'`.
Anyone know how to work around it?
Full message, from `/var/log/zettarepl.log`:
```
[2021/01/16 16:02:40] WARNING [replication_task__task_8] [zettarepl.replication.run] For task 'task_8' at attempt 5 recoverable replication error RecoverableReplicationError("full send of tank/iso@auto-2021-01-16_16-00 estimated size is 21.9G\nfull send of tank/iso@auto-2021-01-16_16-00 estimated size is 21.9G\ncannot unmount '/var/db/system': Device busy\nwarning: cannot send 'tank/iso@auto-2021-01-16_16-00': signal received")
```
As background:
Initially, replication was up and running fine using an earlier FreeNAS release (maybe 18 months ago?). When moving country though, I disabled replication and took the primary data set with me. Now I have a new box for replicating to, and am trying to get replication working again. Updated everything to 11.3-U5, and so far the experience has been terrible. :( :( :(
Due to bug NAS-107821 (not being able to edit existing replication tasks, which should have generated a new FreeNAS point release all for itself!), and this bug with replication failing, I've wiped all the old ('LEGACY') replication tasks, their associated SSH connections, keypairs, and the associated snapshot tasks. Then tried recreating things from scratch.
Setting things up anew (keypair, ssh connection, snapshot tasks, replication schedule) didn't work. With the above error message just happening instead.
Having already looked through the forums, someone else seems to have had the same problem: https://www.truenas.com/community/t...ying-to-do-zfs-replication.85546/#post-592214
However, none of the information there seems to really apply. (?)
I'm trying to replicate an initial very small data set (tank/iso ~20GB) from one box (cluster1) to another box (cluster2). Using a small data set initially, to work out what's needed to make things work. Once it's working, I'll copy the approach and do the other data sets (many TB's).
Anyone have ideas?
I keep on getting an error message about `cannot unmount '/var/db/system'`.
Anyone know how to work around it?
Full message, from `/var/log/zettarepl.log`:
```
[2021/01/16 16:02:40] WARNING [replication_task__task_8] [zettarepl.replication.run] For task 'task_8' at attempt 5 recoverable replication error RecoverableReplicationError("full send of tank/iso@auto-2021-01-16_16-00 estimated size is 21.9G\nfull send of tank/iso@auto-2021-01-16_16-00 estimated size is 21.9G\ncannot unmount '/var/db/system': Device busy\nwarning: cannot send 'tank/iso@auto-2021-01-16_16-00': signal received")
```
As background:
Initially, replication was up and running fine using an earlier FreeNAS release (maybe 18 months ago?). When moving country though, I disabled replication and took the primary data set with me. Now I have a new box for replicating to, and am trying to get replication working again. Updated everything to 11.3-U5, and so far the experience has been terrible. :( :( :(
Due to bug NAS-107821 (not being able to edit existing replication tasks, which should have generated a new FreeNAS point release all for itself!), and this bug with replication failing, I've wiped all the old ('LEGACY') replication tasks, their associated SSH connections, keypairs, and the associated snapshot tasks. Then tried recreating things from scratch.
Setting things up anew (keypair, ssh connection, snapshot tasks, replication schedule) didn't work. With the above error message just happening instead.
Having already looked through the forums, someone else seems to have had the same problem: https://www.truenas.com/community/t...ying-to-do-zfs-replication.85546/#post-592214
However, none of the information there seems to really apply. (?)
I'm trying to replicate an initial very small data set (tank/iso ~20GB) from one box (cluster1) to another box (cluster2). Using a small data set initially, to work out what's needed to make things work. Once it's working, I'll copy the approach and do the other data sets (many TB's).
Anyone have ideas?
