Artix Linux Forum

Are you using nfs-suite66 package?

No I don't. What should I do to configure this init script besides installing nfs-utils-suite66?

I will in a minute. I will edit my reply after I run the commands and reboot the system.

EDIT: Unfortunately, running the above commands did not help.

Yes, I have nfs-utils installed and I just installed nfs-utils-suite66 but I'm a complete newbie to suite66 so I don't know how to set it all up.

EDIT: I can see that nfs-utils-suite66 is for server, not client so I guess I don't need it, right?

In a vm, I used what you had in fstab and had no issue with startup

I added a [stop] part to the mount-fstab script in suite66-scripts-1.7.1.1-1 , currently it in que to build. Once it is built and mirrors syncs, you will get new version. I am unsure this will help or not, but I am hoping it is hanging on fstab part and it will help you.

Thank you. Now I'm thinking maybe the problem is that the network is down before the nfs is unmounted. But this is just my hypothesis...

Trying to order while shutting down is not really possible. At least not with the way s6-rc is designed. Assuming you know what service exactly hangs, a better approach would be to use the finish script to clean it up and make sure it's gone.

Trying to order while shutting down is not really possible. At least not with the way s6-rc is designed. Assuming you know what service exactly hangs, a better approach would be to use the finish script to clean it up and make sure it's gone.

With 66, there is no finish script.

@mattz7 The only thing else I know to try is to edit /usr/bin/reboot and add -f flag.
You can edit /usr/bin/shutdown to do:
66-hpr -pf
I would make a backup of the shutdown script before doing so. The -f flag forces shutdown and only syncs instead of going through normal reboot/shutdown process, so it should be near instant action

Edit: First try Edit 2
Edit 2 : Try changing your rc.shutdown script in /etc/66 to: https://termbin.com/fpmt

So what are you suggesting I should do?

BTW, I'm getting an fstab fatal error now when shutting down (see the picture attached).

I have no further idea on how to solve this :\

Let me update the mount script later when I have time, I have a kid to get ready for school and other stuff to do; and do what s6 does. Which will remove mount-netfs

66-env -t boot -e $EDITOR boot@system and change MNT_NETFS=!yes to MNT_NETFS=!no (where $EDITOR is something like vim ) as root and also run as root: 66-enable -t boot -F boot@system

@mattz7 The only thing else I know to try is to edit /usr/bin/reboot and add -f flag.
You can edit /usr/bin/shutdown to do:
66-hpr -pf

Ok I've removed mount-netfs and changed mount-fstab to only mount all that doesn't have noauto or no_netdev set in mount options. HOPEFULLY this will not mount nfs on boot and no cause issue

Edit: new scripts version is in gremlins

To be clear, if you manually umount the nfs partition before shutting down, does the system still hang?

But the NFS share was not mounted at boot. I had to mount it manually.

Did you manage to read IMPORTANT INFORMATION above? ;) ;) Apparently there is something wrong with the way wifi is handled at shutdown.

Anyhow, I'm waiting for the mirrors to sync and I'm going to try out your new revision from gremlins. Thank you! :)

The system hangs when I poweroff after I do

Code: [Select]

66-stop iwd

  and don't umount the nfs share beforehand.  Only now, I receive one message at shutdown, namely

Code: [Select]

*** WARNING ** The system is going down NOW

Thanks for the all testing! So that does confirm it is the nfs umounting in suite66 causing the hanging. That must mean our patch was not correct. I personally believe this should be handled in the umount step during shutdown in suite66, so I will go look at the code again.

@ndowens: discovered a trivial error (darn C) in the patch so the nfs partition was not being skipped. It's been corrected. The 'suite66-0.6.1.3-2.3' package is in gremlins now so if you could give that a try @mattz7.

Sure I'll give it a try once the mirrors sync, but I also discovered that forcing iwd to umount the nfs share before it stops also solves the issue.

Sure I'll give it a try once the mirrors sync, but I also discovered that forcing iwd to umount the nfs share before it stops also solves the issue.

Thank you!

Of course, I did

Code: [Select]

pacman -S iwd-suite66
66-enable -t default -F iwd 

BTW, do you know which mirror has the 2.3 version you pushed?

EDIT: OK, got it. The last mirror from /etc/pacman.d/mirrorlist.

I will test that in a minute! :)

EDIT2: I had something that read "fstab failed" message, but the system shut down! THANK YOU! You da man! :D

I will mark the thread as solved, but can you further help me troubleshoot the message about fstab?  I will power on the computer and capture a picture of the message.

EDIT3: DAMN! It does not work. Maybe the last time around the iwd service kicked in with the stop procedure.

Edit: for the mount-fstab thing, only happens during shutdown? If so it could be ignored as it shouldn't cause an issue. The command it runs shouldn't cause an issue

Code: [Select]

mount -a -O noauto,no_netdev

OK. I want to thank you @ndowens and @Dudemanguy for your time.

I can't wait to test out iwd-suite66 once it's there!

Hope y'all enjoying your day wherever you are. Greetings from Poland! :)

EDIT:
I had this message the first time around when shutdown went well. I'll test it once again with the modified iwd script.

Dang that's a bummer that did not work. Maybe there's more to nfs mounts than I expected. In the meantime, we'll try to adjust the stop/finish scripts which is worth it anyway. Could you post the output of cat /proc/mounts with your nfs mounted? I'm not sure what nfs mounts look like in that file (so we know what to grep for exactly).

I'm waiting for the mirrors to sync and once everything works, I'm going to mark this thread as solved :)

EDIT: DAMN IT. Now the system hangs. I don't know what happened since the last time. Let me do some more testing.

EDIT2: @ndowens so apparently, I was trying to

Code: [Select]

umount -t nfs -a

while I should be doing

Code: [Select]

umount -t nfs4,nfs -a

The latter option appears to work. Please account for it in the iwd-suite66 package.

BTW, I downgraded suite66 from gremlins to system, and did 

Code: [Select]

MNT_NETFS=!yes

in 66-env -t boot -e $EDITOR boot@system

Thank you! :)

@mattz7 can you edit /etc/66/service/iwd and change umount -a -t .. to umount -f -a -t ... and 66-enable -F iwd and see if that helps

Thanks for the reply :)

Funny thing is it worked now, but let me double check, boot up the system once again, mount the nfs and shutdown again :P

EDIT:
OK, the second time it worked, too.

I'll edit /etc/66/services/iwd and delete the -f option, do 66-enable -F iwd and see if it doesn't work. If it doesn't work, it means you nailed it! :)

EDIT2:

So I reverted the changes in iwd, re-enabled the service, mounted nfs and shutdown the computer. Only this time, it worked too.

Sorry for all this misleading information, but seriously I have such a mindf*** right now. I don't know what has changed since the last time, but apparently now it works.

BTW, do you know why the -f option would make a difference here? Because I don't get it.

I'll check the -f option a couple more times.

EDIT3: So running

Code: [Select]

loginctl poweroff

also causes the system to hang while shutting down.

I always just use terminal to reboot or shutdown. Where did you use the dbus-send cmd?

Side note, I just pushed update elogind-suite66 package that cleans up a bit of it, though I doubt it helps in this situation. I updated mount-cgroups to do partly what elogind script was doing.

I used it in yakuake in KDE plasma ::)

As mentioned above, seems to be an elogind issue. I have never tried shutdown from plasma's menu :) . Usually I am in terminal doing w/e so I just do it there