- Joined
- May 13, 2015
- Messages
- 2,478
I have a small herd? fleet? of FreeNAS servers, most running the ol' reliable FreeNAS version 11.2-U8
Why the older version? Mostly because it 'Just Works' for my simple purposes, but also because I don't like the new user interface -- it makes me work too hard! -- and 11.2-U8 is the last version that lets me use the mono-based 'legacy' GUI.
But also... failover LAGGs are broken in all versions after 11.2-U8: see [NAS-104883] Unable to prioritize interface member in LAGG on the TrueNAS Jira system. A fix has been scheduled for release 13, and this LAGG issue mostly doesn't affect me. I say 'mostly' because all but one of my FreeNAS servers run as VMs under ESXi -- which handily supports failover NIC teaming -- but that one FreeNAS server running on the bare metal is the one server where I use failover LAGG groups.
Lastly, and the reason I created this post: upgrading from 11.2 to 11.3 doesn't break my replication tasks, but upgrading to 12.0 does. I see errors in the debug logs reporting that the remote server can't find
My simple 'fix' is to create a soft link in the /sbin folder to theVoila! My FreeNAS 11.2-U8 replication tasks targeting the TrueNAS 12.0-U5.1 server started working correctly! Hooray!
Now, be warned: this isn't the right way to do this -- you probably ought to just re-create the replication tasks.
But it works. It's quick. It's dirty. It may not survive upgrades. It may even contribute to Global Warming! Who knows? But it suits my purpose in that it let's me boot different FreeNAS/TrueNAS versions on my 'sandbox' system without having to recreate replication tasks every time I switch. And that's good enough for now.
Why the older version? Mostly because it 'Just Works' for my simple purposes, but also because I don't like the new user interface -- it makes me work too hard! -- and 11.2-U8 is the last version that lets me use the mono-based 'legacy' GUI.
But also... failover LAGGs are broken in all versions after 11.2-U8: see [NAS-104883] Unable to prioritize interface member in LAGG on the TrueNAS Jira system. A fix has been scheduled for release 13, and this LAGG issue mostly doesn't affect me. I say 'mostly' because all but one of my FreeNAS servers run as VMs under ESXi -- which handily supports failover NIC teaming -- but that one FreeNAS server running on the bare metal is the one server where I use failover LAGG groups.
Lastly, and the reason I created this post: upgrading from 11.2 to 11.3 doesn't break my replication tasks, but upgrading to 12.0 does. I see errors in the debug logs reporting that the remote server can't find
/sbin/zfs
. This made sense after I snooped around on my new 12.0 server and found that zfs
has 'moved' from the /sbin directory to /usr/local/sbin.My simple 'fix' is to create a soft link in the /sbin folder to the
zfs
program in /usr/local/sbin, like this:Code:
ln -s /usr/local/sbin/zfs /sbin/zfs
Now, be warned: this isn't the right way to do this -- you probably ought to just re-create the replication tasks.
But it works. It's quick. It's dirty. It may not survive upgrades. It may even contribute to Global Warming! Who knows? But it suits my purpose in that it let's me boot different FreeNAS/TrueNAS versions on my 'sandbox' system without having to recreate replication tasks every time I switch. And that's good enough for now.