On 19.07.2024 16:28, mh@mike.franken.de wrote:
On Freitag, 19. Juli 2024 15:21:23 MESZ Andrei Borzenkov wrote:
On Fri, Jul 19, 2024 at 3:57 PM <mh@mike.franken.de> wrote:
On Samstag, 13. Juli 2024 16:11:41 MESZ mh@mike.franken.de wrote:
Hi *,
starting with the last two NetworkManager versions, my dispatcher script does not finish when going to sleep. It finishes on resume, though, causing some strange effects. It's runtime is 1 or 2 seconds only, so this is not a problem of hitting the dispatcher timeout. Anyway in this case it wouldn't continue after resume, I believe.
Any idea, what is happening here, and how to avoid it?
It seems, that NM takes down the dhcp lease, even before the dispatcher service starts. So there is no chance to umount nfs shares due to a lack of connectivity. Assuming you are talking about the standard nfs script which is part of openSUSE NetworkManager package - the very idea of this script is to unmount when the NFS server is no longer accessible.
No, I wrote my own script, because the default script wasn't sufficient.
Without seeing this script everything will be just a wild guess.
And as already said - it worked for years until NM 1.46.
Did you verify that downgrading NetworkManager (and *only* NetworkManager) restores the previous behavior?
And yes, in the past my experience with NFS was - if the server does not answer, attempt to unmount hangs. One hack was to configure the server IP on a local interface, then "umount -l -f" would succeed.
Thats what my script already does.
Thx anyway.