Re: [PVE-User] Fwd: Replication Job: 101-0 failed

Re: [PVE-User] Fwd: Replication Job: 101-0 failed

[Marco Gaiarin]

Mandi! Julival Santos Reis Júnior
  In chel di` si favelave...

> I'm having this same problem.
> Did you manage to solve it, or discover something?

Ops, forgot to report on list.

The message seems is generated by 'overlapping' replication tasks, eg when a
replication task does not end before the next one start. And so, is totally
harmless.


Also, but i have to test better, performance of replication can be inproved
simply by increasing the 'blocksize' parameters in the ZFS storage
configuration.

AFAI've understood, the 'blocksize' is the 'transfer unit' of a ZFS
replication operation, and so augmenting it (i'm testing 32k) can improve
performance.

Re: [PVE-User] Fwd: Replication Job: 101-0 failed

[Marco Gaiarin]

Mandi! Julival Santos Reis Júnior
  In chel di` si favelave...

> Thanks for the feedback!

Please, don't reply personally, use the list!!!


> I had solved this problem in version 6 of proxmox.
> There was a parameter that I increased the timeout, and the system stopped
> giving these errors. But now it changed to version 7 and the problem came back.
> sending a print of the file in version 7, from the place I moved. I also read
> that the problem is solved by putting good quality SSD enterprise hard drives.

'timeout' isthe simplier consequence of other things, typically
load/iodelay.
So augmenting the timeout you are only curing the symptom, not the disease.

Try increase the blocksize (or, sure, spend more money and by good
enterprise SSD drives! ;-), instead of modifying the code for increase the
timeout...