Has the uneven recordsize / replication bug been fixed?

TrueNAS General
, , ,
1

Back in 2020, a bug was reported on the forum here regarding the issues that can arise out of replicating a dataset with one recordsize to a remote server where the recordsize for a given dataset was different. Does the Replication task still have to feature a “–large” flag as described here, or is TrueNAS smart enough to know how to send larger datasets with large recordsizes w/o error?

If the –large flag has to be set, how would I go about it since I see no mention in the Replication task GUI pane to do so?

2

iirc, yes; I would have to dig for something to link you, but I don’t have the energy to do so.

3

No worries. With these reports on the internet, it’s hard to know what has and has not been resolved, especially if the bug was reported in 2020.

I still don’t see how you can pass flags within the TrueNAS GUI to the replication process though. So since I played around with record sizes on this end I decided to play it safe, nuked the receiving remote from orbit and started the replication from scratch. It’ll take a while but it should also be bugproof since all recordsizes, sVDEV parameters, etc will match.

I hope iXsystems can chime in at some point re: flag use and / or the bug itself, however.

4

You don’t have this option?

large-blocks

5

I didn’t see it because I was looking for a spot to put the --large flag in (sort of like the vfs fruit, etc. back in the more complicated days of supporting time machine) but you are right, the GUI has the feature as a one-time use check-box. Luckily mine was already checked!

Thank you for that!