At high level,
a) some process is mounting an image file in UDF format into an in memory virtual disk and failing to allocate sufficient memory to complete.
b) the open source nvidia driver or the waylaid process is then faulting because one of its page files is missing.
My working guess is : and its pure guess…
The nouveau driver or the waylaid press on top isn’t memory safe. so it’s causing the hang while it's waiting for a memory segment that’s been stollen by the loop memory allocation process.
(Being gpu that’s probably why its only impacting gui threads e,g, chrome and not ssh)
The real bug is why did the loop a) steal the memory? It should have been marked as protected/reserved ? Theorising that Wayland is likely source of the problem… probably not checking/allocating properly.
On 10 Sep 2025, at 09:57, rmluglist2--- via Hampshire <hampshire@???> wrote:
Hi all
Can anyone help me understand what’s going on with a seemingly random (but happening more and more) Ubuntu Jammy freeze please? A copy of kern.log (I’ve not found anything untoward in syslog) that happened around the time of the freeze is given below:
Sep 10 07:21:24 gower kernel: [ 160.289985] loop47: detected capacity change from 0 to 16156496
Sep 10 07:21:25 gower kernel: [ 160.616312] UDF-fs: warning (device loop47): udf_load_vrs: No anchor found
Sep 10 07:21:25 gower kernel: [ 160.616319] UDF-fs: Scanning with blocksize 512 failed
Sep 10 07:21:25 gower kernel: [ 160.616583] UDF-fs: warning (device loop47): udf_load_vrs: No anchor found
Sep 10 07:21:25 gower kernel: [ 160.616587] UDF-fs: Scanning with blocksize 1024 failed
Sep 10 07:21:25 gower kernel: [ 160.616776] UDF-fs: INFO Mounting volume 'VCD0385', timestamp 2005/05/27 10:19 (1000)
Sep 10 07:21:38 gower kernel: [ 174.091545] nouveau 0000:10:00.0: fifo: fault 00 [READ] at 0000000000000000 engine 00 [PGRAPH] client 05 [
HUB/CTXCTL] reason 02 [PAGE_NOT_PRESENT] on channel 1 [003fc46000 Xwayland[4808]]
To explain a bit, this is my homelab fileserver and it works (in terms of being a server) 24x7x365. I also use it to do the odd google, play the odd mp3 / spotify while I’m at work on a completely separate machine. It’s when I do pretty much anything on the machine itself (open Chrome, find a file in nautilus) that it freezes completely. So hard does it freeze that no amount of keystrokes can unfreeze it – it has to be a hard reset using the machine’s reset switch. Curiously, while it’s completely frozen, it is still possible to ssh into it – from where the log file was taken. Even a sudo reboot from the ssh session doesn’t reboot the box, it just disconnects and prevents any further connections.
The only thing I can offer is as follows. The machine has 3*5-bay JBOD USB devices attached i.e. 15 drives in 3 units. I have recently switched one of these off though the fstab entries for the drives in said JBOD haven’t been commented out. As they’re USB and / isn’t mounted on any of them, it’s been working fine. I can, of course, comment the drives in this JBOD out in fstab but, as I say, it has been working fine and, as we know, it’s a little bit trial and error to identify which drives in this particular jbod relate to which mount points in fstab. My understanding, which may well be wrong, is that if the kernel can’t find a device for the UUID in fstab, it just carries on to the next one.
To me, it doesn’t seem likely this is the culprit and I’ve seldom seen an Ubuntu box freeze so badly and still allow ssh sessions. Any ideas anyone? I’ve done a lot of googling but this seems to be a difficult thing to google for.
Cheers
RM
--
Please post to: Hampshire@???<
mailto:Hampshire@mailman.lug.org.uk>
Manage subscription:
https://mailman.lug.org.uk/mailman/listinfo/hampshire
LUG website:
http://www.hantslug.org.uk<
http://www.hantslug.org.uk/>
--------------------------------------------------------------
--
Please post to: Hampshire@???
Manage subscription:
https://mailman.lug.org.uk/mailman/listinfo/hampshire
LUG website:
http://www.hantslug.org.uk
--------------------------------------------------------------
