VLC stopping playback when jumping to timestamps for media on SMB share

TrueNAS General
,
21

I have connected my client directly over ethernet to the management interface of the new box.
I’m getting yet another new hash when testing over SMB:

6039E75D34A576FCEAC6247CC74ECAE7FA6A654F

I will try another client, but I really can’t understand what this is telling me is a problem now?

EDIT: The has keeps changing!
Without doing anything, if I run the hash again, it is different to the one I literally got 5 minutes ago:

3B4EC21A4B8E065303E8BC1EFD3198AF01919592
22

If you’re running 24.04.2.4 or 24.10.0.1, you should immediately upgrade to 24.04.2.5 or 24.10.0.2, there’s an SMB AIO read-buffer issue

2 Likes
23

Ahha! I am indeed on ElectricEel-24.10.0.1

I had only just upgraded before I started the migration, so didn’t even think to check for updates or the release notes d’oh.

I will update now and feedback on the outcome! Tyvm

1 Like
24

Ding ding ding!

Looks like that’s done the trick :smiley:

30F9CA066F5F62F22BA100085640F50F64B48945

Thanks very much for everyone’s help with this, greatly appreciated!

2 Likes
25

That’s crazy! :astonished:

Glad the mystery is finally resolved.

But now I must wonder, @eds89: This could implicate files that you had already transferred over SMB when on SCALE 24.10.1.

@HoneyBadger Is this related to SMB + Linux/kernel, or upstream SMB in general? Would this also include TrueNAS Core 13.3?

26

Doesn’t impact CORE, and it’s only on the read side, so any writes done over SMB will still be correct on-disk.

1 Like
27

I guess to be more accurate, what I meant was this:

Reads from the server to be written to a client’s non-ZFS storage.

Which is why I’m wondering about this:

Because as seen in an earlier post, @eds89 ended up with a corrupted file on the client PC when it was copied from the SCALE server. :warning:

Which means that without confirming this corruption via SHA1 hash comparison, it’s possible that other files (copied from the server to the client) were also corrupted, without manually confirming their integrity.

28

In my case, I have replicated the data again using ZFS replication, and it has NOT yet been copied off that new server.

If the issue is only on the read side, I’m happy with the integrity of the replicated data, and now I’ve completed the update to fix the read issue, I can happily copy files off there now without too much concern of corruption :+1:

Appreciate the line of thinking though

2 Likes