| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Constant SELinux errors from BOINC and WCG. |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 16
|
|
| Author |
|
|
JoeZeff
Cruncher Joined: Dec 31, 2004 Post Count: 33 Status: Offline Project Badges: 
|
I've been running BOINC on a computer running Fedora Linux for several years now. Currently, it's Fedora 14, fully updated. As of this morning, I've been getting tons of SELinux errors from my current WCG unit, wcg_dsfl_6.19_i686-pc-linux-gnu. I've tried following the instructions in the troubleshooter for creating a new policy for them, but it Just Doesn't Work. For more details, here's a link to the bug report at Bugzilla: https://bugzilla.redhat.com/process_bug.cgi
|
||
|
|
Falconet
Master Cruncher Portugal Joined: Mar 9, 2009 Post Count: 3265 Status: Offline Project Badges: 
|
---------------------------------------- AMD Ryzen 5 1600AF 4C/8T 3.2 GHz - 85W AMD Ryzen 5 2500U 4C/8T 2.0 GHz - 28W Intel Z3740 4C/4T 1.8 GHz - 6W |
||
|
|
JoeZeff
Cruncher Joined: Dec 31, 2004 Post Count: 33 Status: Offline Project Badges:
|
That seems to be the issue, and I've reported that to Bugzilla. (They're also looking into a fix from the SELunx side.) However, I'm still getting the alerts. How do I tell BOINC to abort the unit and download a new one?
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I've been running BOINC on a computer running Fedora Linux for several years now. Currently, it's Fedora 14, fully updated. As of this morning, I've been getting tons of SELinux errors from my current WCG unit, wcg_dsfl_6.19_i686-pc-linux-gnu. I've tried following the instructions in the troubleshooter for creating a new policy for them, but it Just Doesn't Work. For more details, here's a link to the bug report at Bugzilla: https://bugzilla.redhat.com/process_bug.cgi Hi JoeZeff, Your link is only giving a log in screen, not the bug report. Can you provide a direct link, or paste a copy here, so we can get the gist. From the errors, have any DSFL result been reported and if so, can you go to My Grid > Result Status and click on a Error link in the Status column to see what it says there. (please copy here). SELinux [Security Enhanced Linux] was not known to me. Were you computing before at WCG with that machine? Did you try running CEP2 (Clean Energy) on this machine? Just curious, because DSFL and CEP2 are slightly different, having a control app and a worker app. Maybe other members can chip in who're familiar with the SELinux feature. ttyl --//-- |
||
|
|
JoeZeff
Cruncher Joined: Dec 31, 2004 Post Count: 33 Status: Offline Project Badges:
|
Sorry, the proper url is https://bugzilla.redhat.com/show_bug.cgi?id=735648
I've been running BOINC for the WCG and Einstein@home on this same machine running Fedora Linux (with SELinux) for about four years now. The Einstein@home units give a little SELinux trouble, but correcting the issue once per unit is enough. I've not finished any DSFL units as yet, so I can't report on any errors. As far as CEP2 goes, I've not had any work units on it as yet, but it's selected. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
For reference, this seems to have been addressed in https://secure.worldcommunitygrid.org/forums/wcg/viewthread_thread,31690
I was having the same trouble. I aborted the three DSFL units, and things have cleared up. |
||
|
|
JoeZeff
Cruncher Joined: Dec 31, 2004 Post Count: 33 Status: Offline Project Badges:
|
Yes. That's exactly the same thread you referenced earlier. How do I abort a unit if this comes up again?
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I'm afraid I continue to see SELinux denials of wcg_dsfl_6.19. From the log messages, it appears to be attempting to walk the /proc file system, which it should not be doing. It's OK (allowed and no error) to examine the tree beneath its own process ID, but it should stay out of the other process ID directories.
When I notice one of these misbehaving, I start the BOINC Manager, select Tasks, select the problematic job, then click Abort in the left hand panel. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The errors continue. I normally participate in all projects, and let WCG choose which are important, but I want to opt out of this one, since it is still broken - it conitnues to try to walk the /proc tree which is a security violation, but might not be flagged in the absence of SELinux. Is there a simple way to opt out of this single project?
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Yes, go to the My Grid > My Projects to (de)select any science you want / don't want. The selection in My Projects will apply to ALL device profiles until a device profile is revised directly. For a specific device, associate this to a specific profile (You can create home, work, school as locations in addition to default)
--//-- |
||
|
|
|