Replication to Core failed due to SSH timeouts

[aadje93]

Hi truenas Team,

after running my 'main' (better call it "active") truenas system on scale, i noticed that randomly the replication fails or succeeds. When it fails its due to SSH timeouts. Is there a setting i can change on either scale or core (the backup machine/replication target) to stop ssh from timing out? Or maybe in the replication settings a change i should do on scale's end?

I'm running a full replication of snapshots to the from scale (RC2 and now release) to core, but every now and then i see notifications for a failed replication. On core -> core i didn't had this and all seemed well.

Do note, i'm daily replicating a couple TB of changed data.... So not your typical home NAS :)

 

[morganL]

If you can confirm the network is reliable, i'd suggest reporting this as a bug when it next happens. It would be useful to capture debugs on both machines.

You might also check ClientAliveInterval settings for ssh - it should be 3 by default. If you are replicating via an unreliable network (e.g Internet) you might want to increase that. https://www.truenas.com/docs/scale/systemsettings/servicesscale/

There's also a ClientAliveCountMax that can be increased

 

[aadje93]

Network is defintly stable as the issue arised right after the core -> scale upgrade on the main machine. Statping shows a <0.01% packet loss to google.com so it seems all is working fine. The replications happens on the same switch both machines are attached to, so that shouldn't be an issue.

I'll look after it and make bug report next time. Does the debug dump get handled securely?

 

[morganL]

I'll look after it and make bug report next time. Does the debug dump get handled securely?

Its not publicly available, but iX engineers will need to review. Let the eng team know any concerns.