23.10.1 - replication direction reversed?

chuck32

Guru
Joined
Jan 14, 2023
Messages
623
Hello all,

Source System
TrueNAS-SCALE-23.10.1
Supermicro X10SRi-F
Xeon 2640v4
128 Gb ECC RAM
Target System
TrueNAS-13.0-U6.1
Supermicro X10SLL-F
i3 4130
16 Gb ECC RAM

I setup a second server with some new hardware and some hardware I had laying around as a target for my replication tasks. I followed this post in order to setup the SSH connection.
First question here would be: I was not prompted for my root password yet. From what I gathered Use Sudo for ZFS Commands should be checked if I have not setup passwordless sudo. I did not change the configuration to the default core root user yet. Is this correct?

Second question:
1703280104051.png

This is on the source machine, neptune is the source pool and earth is the remote target pool.

1703280308122.png

The last part in red confuses me: no snapshops of earth/moon would be replicated. But I do not want to replicate from there, I want to replicate to there. If I change the direction to pull, the message disappears.

Also, is there any advantage using the advanced options over the wizard? I just want to create a 1:1 copy of my main system.

Unfortunately I couldn't test the replication using the wizard because I got the error:
Replication "neptune/test-dataset - earth/moon" failed: Sorry, user root is not allowed to execute '/usr/local/sbin/zfs get -H -p -t filesystem,volume type earth/moon' as root on truenas-core.local..
When I tried to add /usr/local/sbin/zfs for the root user on my scale system it said it couldn't be found.

I tried looking at the documentation but I found it confusing, especially the part about the home directories. Why would I need a home dir for my users?
If using a TrueNAS CORE system as the remote server, the remote user is always root.
But it seems like the configuration has to be made for the local scale users?

It seems like I'm misunderstanding some crucial part here. I thought CORE as the remote system would have some advantages (lower CPU usage because the lacking apps / VM capabilities, higher memory usage for cache (although that shouldn't really matter I guess; I can only replicate over power lan, which will limit me to 5 Mib/s or something like that).
I might consider switching to SCALE though because if I remember correctly, when I played around with a SCALE test system the remote replication worked basically out of the box.
 
Last edited:

chuck32

Guru
Joined
Jan 14, 2023
Messages
623
Update

I switched over to SCALE for the replication and encountered some similiar troubles, which I was able to resolve though.

1) The red warning message, which seems to indicate a reversed direction persisted
2) I wasn't able to successfully replicate until I added newpool/exisitingdatasetname manually in the destination field. I will still receive the message that an error occured because the dataset does not exist, but the replication task will then continue normally. The source datasets are encrypted.

Remaing questions:

Is the behaviour regarding the error message intended or should I file a report?

Do I need to remove full filesystem replication after the initial sync?

When trying to allow pwless sudo for /var/sbin/zfs I receive an error:
1703367196629.png

I'd like to avoid allowing all sudo commands with no password if possible.

On a side note to myself: find an alternative to powerline LAN, it's not even faster than uploading to the cloud... Only advantage is the second system is not in my appartement and since that room is not heated I don't need to worry about temps and noise.

Thank you all and happy holidays!
 

morganL

Captain Morgan
Administrator
Moderator
iXsystems
Joined
Mar 10, 2018
Messages
2,694
Update

I switched over to SCALE for the replication and encountered some similiar troubles, which I was able to resolve though.

1) The red warning message, which seems to indicate a reversed direction persisted
2) I wasn't able to successfully replicate until I added newpool/exisitingdatasetname manually in the destination field. I will still receive the message that an error occured because the dataset does not exist, but the replication task will then continue normally. The source datasets are encrypted.

Remaing questions:

Is the behaviour regarding the error message intended or should I file a report?

Do I need to remove full filesystem replication after the initial sync?

When trying to allow pwless sudo for /var/sbin/zfs I receive an error:
View attachment 73786
I'd like to avoid allowing all sudo commands with no password if possible.

On a side note to myself: find an alternative to powerline LAN, it's not even faster than uploading to the cloud... Only advantage is the second system is not in my appartement and since that room is not heated I don't need to worry about temps and noise.

Thank you all and happy holidays!
TrueCommand 3.0 has a replication wizard for these types of situations. It simplifies the security requirements.

 
Top