https://bugzilla.novell.com/show_bug.cgi?id=758379 https://bugzilla.novell.com/show_bug.cgi?id=758379#c1 --- Comment #1 from carl mcgrath 2012-04-20 23:50:57 UTC --- Created an attachment (id=487316) --> (http://bugzilla.novell.com/attachment.cgi?id=487316) segment of /var/log/messages where STR fails -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=758379 https://bugzilla.novell.com/show_bug.cgi?id=758379#c2 carl mcgrath changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |cmcgrath5035@comcast.net --- Comment #2 from carl mcgrath 2012-04-22 11:32:13 UTC --- Issue unchanged (still a bug) after recent update to kernel 3.1.10-1.9. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=758379 https://bugzilla.novell.com/show_bug.cgi?id=758379#c4 --- Comment #4 from carl mcgrath 2012-04-30 16:57:13 UTC --- I understand your analysis. I originally opened https://bugzilla.novell.com/show_bug.cgi?id=751811, when STR stopped working in the KDE 4.7.x to 4.8.1 transition. That bug was fixed for most people, but I was left with the residual NFS issue. I can only assume some change in NM between KDE 4.7 and 4.8 is responsible, but that is a large space. For now I have added an umount (on suspend) and mount(on restore) to /etc/pm/sleep.d which is working OK for me. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=758379 https://bugzilla.novell.com/show_bug.cgi?id=758379#c6 --- Comment #6 from carl mcgrath 2012-05-02 11:13:40 UTC --- My temporary solution is a script in /etc/pm/sleep.d/00aa_cjm_nfsworkaround. It runs first on suspend, last on resume, I can see it's results in the /var/log/pm-suspend.log file. On suspend, it calls another script, umount_nfs, that umounts each of my nfs folders. On resume, it calls another script, mount_nfs, that mounts those folders. The go-into-suspend works fine, and the system sleeps. The resume-from-suspend does not work, although the machine comes back to proper working order. Each mount call in the mount_ script fails, with a message in /var/log/pm-suspend.log. I am suspecting that although NM is back up and the physical network interfaces have been brought up by /user/lib/pm-utils/sleep.d/0055NetworkManager, the DHCP processes may still be running and the network not ready for the mount calls. Manually running the mount_nfs script from a konsole works fine and the NFS shares are then restored, after a human response time, of course. I am going to experiment with a delay in the restore section of my 00aa_cjm_nfsworkaround script. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=758379 https://bugzilla.novell.com/show_bug.cgi?id=758379#c8 --- Comment #8 from carl mcgrath 2012-05-04 03:57:57 UTC --- I had removed my NFS mounts from fstab as part of my work around strategy, but realized that might not be necessary. I put them back in, and now the mounts are reestablished on resume. For the moment, I'll say it is done by the /etc/NetworkManager/dispatcher.d/nfs wlan0 up script. I'll dig deeper tomorrow -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=758379 https://bugzilla.novell.com/show_bug.cgi?id=758379#c10 --- Comment #10 from carl mcgrath 2012-05-08 13:53:55 UTC --- Resume after a 48 hour sleep worked OK, so I'll say I cannot reproduce what I thought was a DNS related error. I do know that one of the two "by-name" servers was down at the time, unsure why that would cause both to fail re-mount. After a recent update from KDE 4.8.2 to KDE 4.8.3, no change, I still have to un-mount NFS shares for sleep to complete. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=758379 https://bugzilla.novell.com/show_bug.cgi?id=758379#c12 --- Comment #12 from carl mcgrath 2012-05-31 13:02:01 UTC --- Summary of recent experiments: Works: 1. Execute "sudo umount -alt nfs,nfs4" from konsole cli. 2. Initiate sleep from klauncher - shutdown-sleep 3. On wake, nfs folders are remounted Does not work: 1. Install script file in /etc/pm/sleep.d (see attached) 2. Initiate sleep from klauncher - shutdown-sleep The script does run first, returns success (as seen in /var/log/pm-suspend.log), but krunner and plasma-desktop refuse to Freeze, Trace in /var/log/messages shows waiting on nfs related stats. Works: 1. Execute "sudo umount -alt nfs,nfs4" from konsole cli. 2. Initiate sleep from klauncher - shutdown-sleep, with the script still installed. For some reason, running the script does not work, although separately running the same command first does. I tried adding delay after the umount in the script (up to 7 seconds), not a fix. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=758379 https://bugzilla.novell.com/show_bug.cgi?id=758379#c14 --- Comment #14 from carl mcgrath 2012-06-01 13:14:21 UTC --- Gary - I think I now understand the issue - please comment: I was thinking thru your recent comment about moving the umount to 54, wondering why you thought that might help. I then began to wonder if the script was really running successfully, in spite of it reporting success in /var/log/pm-suspend.log. I was formulating a way to add diagnostics to the script to check if it was in fact running, but I then realized, and now am thinking, that when I initiate sleep as a user (not as root), somewhere there is a setup that allows users to initiate sleep but, perhaps, the scripts run then as user, not as root, so therefore umount in a script may not really work. So, I quickly tried from a konsole cli "sudo /usr/bin/pm-suspend" , I am prompted for root password as I would expect (due to sudo) and the system sleeps as desired. I also confirmed that WITHOUT my 00script, "sudo /usr/bin/pm-suspend" does not work. Seems to confirm to me my suspicion that 00script is not really working correctly when it runs without sudo, presummably as user. Comments? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=758379 https://bugzilla.novell.com/show_bug.cgi?id=758379#c16 Aaron Digulla changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |digulla@hepe.com --- Comment #16 from Aaron Digulla 2012-06-29 19:36:15 UTC --- I seem to have the same issue. I upgraded KDE from 4.7.x (what you can find in the official openSUSE 12.1 release) to the version that can be found in this repo: http://download.opensuse.org/repositories/KDE:/Extra/KDE_Release_48_openSUSE... After the update, suspend to disk didn't work anymore. The error message is exactly the same that you get in the attached /var/log/messages (plasma-desktop and krunner hang in rpc_wait_bit_killable). I'm *not* using NetworkManager. Unmounting NFS solves the issue for me as well. What I don't understand: 1. Why is KDE 4.8 affected but 4.7 isn't? 2. Why is krunner and plasma-desktop using NFS?? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.