Proof of time chain reorgs drop farmers out of sync for 1s

Support
1

A proof of time chain reorg seems to cause farmers to claim the node is out of sync for 1s

Node log:

2023-11-17T05:39:13.710077Z [Consensus] Proof of time chain reorg happened from_next_slot=1426692 to_next_slot=1426692

Farmer 1 log:

2023-11-17T05:39:14.052697Z  INFO single_disk_farm{disk_farm_index=0}: subspace_farmer::single_disk_farm::plotting: Node is not synced yet, pausing plotting until sync status changes
2023-11-17T05:39:15.052278Z  INFO single_disk_farm{disk_farm_index=0}: subspace_farmer::single_disk_farm::plotting: Node is synced, resuming plotting

Farmer 2 log:

2023-11-17T05:39:14.620388Z  INFO single_disk_farm{disk_farm_index=0}: subspace_farmer::single_disk_farm::plotting: Node is not synced yet, pausing plotting until sync status changes
2023-11-17T05:39:15.620461Z  INFO single_disk_farm{disk_farm_index=0}: subspace_farmer::single_disk_farm::plotting: Node is synced, resuming plotting

Happens occasionally.

Not sure if that is a real issue or just too extensive logging on the farmer side.
I just thought I report the observation.

2

Can you provide a bit more logs from the node, like 10-20 lines before and after that line?

In general PoT chain reorg never happens during normal operation, it only primarily happens after chain sync, which likely means your node did get out of sync and then it synced, which in turn resulted in PoT chain reorg. So it is PoT chain reorg that happens because you were out of sync, not the other way around.

3

Output of grep -B200 -A100 05:39:13.710077 node.log (timestamp of resync)

Is there a way to tell the node is out of sync? Long Idle cycle for sure here.

Node has been running 8 hours when reorg happened.

4

Looks like there was networking glitch either on your Internet connection or in your region more broadly. I see blocks imported by my node while yours didn’t. You can also see less connections right after it recovered, meaning there was probably connectivity outage and some connections expired.

So far I do not think it is an implementation issue. Try running ping to some known good server and see if there is any correlation in delays/dropped packets that correlate with node de-sync.