Message boards : BOINC client : Recent boinc client problem in Fedora 16
Message board moderation
Previous · 1 · 2
Author | Message |
---|---|
Send message Joined: 1 Jan 08 Posts: 18 ![]() |
Here is what happened: [root@semazen hako]# /sbin/chkconfig boinc-client off [root@semazen hako]# ---Reboot--- [hako@semazen ~]$ ps aux |grep boinc hako 2401 0.0 0.0 109240 876 pts/1 S+ 13:13 0:00 grep --color=auto boinc [hako@semazen ~]$ [hako@semazen ~]$ mkdir clienttest [hako@semazen ~]$ cd clienttest/ [hako@semazen clienttest]$ boinc_client 27-Dec-2011 13:16:03 [---] Starting BOINC client version 6.12.35 for x86_64-pc-linux-gnu 27-Dec-2011 13:16:03 [---] log flags: file_xfer, sched_ops, task 27-Dec-2011 13:16:03 [---] Libraries: libcurl/7.21.7 NSS/3.12.10.0 zlib/1.2.5 libidn/1.22 libssh2/1.2.7 27-Dec-2011 13:16:03 [---] Data directory: /home/hako/clienttest 27-Dec-2011 13:16:03 [---] Processor: 8 GenuineIntel Intel(R) Core(TM) i7-2600K CPU @ 3.40GHz [Family 6 Model 42 Stepping 7] 27-Dec-2011 13:16:03 [---] Processor: 8.00 MB cache 27-Dec-2011 13:16:03 [---] Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc aperfmperf pni pclmulqdq dtes64 mo 27-Dec-2011 13:16:03 [---] OS: Linux: 3.1.6-1.fc16.x86_64 27-Dec-2011 13:16:03 [---] Memory: 7.78 GB physical, 2.93 GB virtual 27-Dec-2011 13:16:03 [---] Disk: 240.31 GB total, 99.40 GB free 27-Dec-2011 13:16:03 [---] Local time is UTC -6 hours 27-Dec-2011 13:16:03 [---] NVIDIA GPU 0: GeForce GTX 580 (driver version unknown, CUDA version 4010, compute capability 2.0, 1536MB, 1581 GFLOPS peak) 27-Dec-2011 13:16:03 [---] No general preferences found - using BOINC defaults 27-Dec-2011 13:16:03 [---] Preferences: 27-Dec-2011 13:16:03 [---] max memory usage when active: 3983.57MB 27-Dec-2011 13:16:03 [---] max memory usage when idle: 7170.43MB 27-Dec-2011 13:16:03 [---] max disk usage: 10.00GB 27-Dec-2011 13:16:03 [---] don't use GPU while active 27-Dec-2011 13:16:03 [---] suspend work if non-BOINC CPU load exceeds 25 % 27-Dec-2011 13:16:03 [---] (to change preferences, visit the web site of an attached project, or select Preferences in the Manager) 27-Dec-2011 13:16:03 [---] Not using a proxy 27-Dec-2011 13:16:03 [---] This computer is not attached to any projects 27-Dec-2011 13:16:03 [---] Visit http://boinc.berkeley.edu for instructions Initialization completed Looks promising to me. I didn't do anything more as per your recommendation :). I will wait for further recommendations, but may I suggest to clean all /var/lib/boinc directory before doing anything else since there is some stuff there accumulated during our tests, etc.? |
![]() ![]() Send message Joined: 6 May 06 Posts: 287 ![]() |
Here is what happened: As I said previously If you get that far join a new project and let it crunch one wu as a test. Sign up at another project other than climate prediction. You should be able to open the boinc manager (gui) and it connects to your running session in clienttest. You can check this from in the manager by checking the event log (Advanced Menu/Event Log or Shift Ctrl E) and check that event log is identical to the output you posted from the terminal. Leave /var/lib/boinc for the time being this instance of boinc is not using it - its using 27-Dec-2011 13:16:03 [---] Data directory: /home/hako/clienttest What we are doing is finding out if there is anything wrong with boinc itself, once we have that sorted then we can figure out how to get you back running as a service/daemon. CIC1=CC=C(C2=N[C@@H](CC(OC(C)(C)C)=O)C3=NN=C(C)N3C4=C2C(C)=C(C)S4)C=C1 |
Send message Joined: 1 Jan 08 Posts: 18 ![]() |
... I can start the boinc manager with the 'boincmgr' command fron the clienttest directory but it does not connect to the client. I tried using the select computer > localhost option and pasted the contents of ~/clienttest/gui_rpc_auth.cfg file in the password field but it didn't work. Since I cannot connect to the client I cannot start add project/account manager wizard.
Understood. |
Send message Joined: 1 Jan 08 Posts: 18 ![]() |
boinc_gui from the clienttest directory works. I will post again once the project starts and I can view the event log. |
![]() Send message Joined: 4 May 09 Posts: 8 ![]() |
It seems like it did the trick Dagorath. I followed your instructions [snip] Hmmmm... how did you follow the instructions when in step 4 it says to 'cd to /usr/bin/boinc' ??? i.e. There is no /usr/bin/boinc subdir. |
Send message Joined: 1 Jan 08 Posts: 18 ![]() |
@Trog Dog: Seti at Home runs well by user hako from /clienttest directory at this time. I believe it may support the point that the problem is a wu/app problem. Do you think I am ready to test the daemon service test at this point? If so, please advise. @Darr247: You caught the mistake well. I was actually issuing the command from /usr/bin at that point. |
Send message Joined: 1 Jan 08 Posts: 18 ![]() |
Well, I still have the ghost process after starting the boinc service as a daemon. The inability to launch boinc at boot time therefore may not be a wu/app problem. A reply to my post in fedora forum (http://forums.fedoraforum.org/showthread.php?t=274341) suggested that this may be due to an update on SElinux. I don't have the guts to install a testing version of SElinux at this time. If there is another way to test this hypothesis I may do that, but otherwise I guess I will wait for the next stable update. In the mean time how about an ugly workaround like the following? /sbin/chkconfig boinc-init onwhere boinc-init is a bash script something like #!/bin/bash cd /home/hako/clienttest su -l hako -c 'nohup nice -19 /usr/bin/boinc_client' exit 0 |
![]() ![]() Send message Joined: 6 May 06 Posts: 287 ![]() |
erm, actually no I was going to suggest killing the clienttest instance by doing ctrl - C in that terminal then restarting boinc running against your /var/lib/boinc by issuing the command boinc_client --dir /var/lib/boincand seeing the outcome - that would test whether your wu's were the cause. but at this point I think it's time to start over :( use your package manager to uninstall boinc, delete /var/lib/boinc and boink, delete the directory you downloaded the berkeley package into, reboot, do a locate boinc (as root) and delete any files it lists, do a whereis boinc (as root) and delete any files it lists, reboot - confirm there are no traces of boinc on your hard disk then reinstall boinc from your package manager. CIC1=CC=C(C2=N[C@@H](CC(OC(C)(C)C)=O)C3=NN=C(C)N3C4=C2C(C)=C(C)S4)C=C1 |
Send message Joined: 28 Dec 11 Posts: 3 ![]() |
Hopefully I have something of value I can contribute in this thread. I too have been encountering difficulties with the recent boinc-client update in the fedora 16 distribution. First, after upgrading to 6.12.35, I was no longer able to connect to localhost with the Boinc Manager, however the client continued run, uploading and downloading work etc. I decided yesterday to "dig in" once more to see if I could figure something out. I found this thread and to my dismay, I also noticed that the client was no longer running on my system and I now had this ghost boinc_client process as well. After some more digging, I went down the path of selinux having something to do with this issue, and performed the following command. grep "denied" /var/log/audit/audit.log In my case, there were numerous entries indicating the boinc_client had been denied. Additionally,in my /var/log/messages file, there were numerous entries with regards to the boinc_client as well. I set selinux to disabled and rebooted the system and then was able to start the boinc-client successfully has well as getting the boinc manager to connect to the localhost! I then set the selinux policy back to enforcing and rebooted. Some more poking around led me to some open bugs with regards to boinc and selinux and on a hunch, I went into the updates-testing repository and found new selinux policy updates which I applied. 3.10.0-69.fc16 (Sorry, but I didn't record the selinux bug numbers I found) I then started the client and now, no ghost process! Then I started the manager and was able to successfully attach to the localhost as well! Perhaps this something which should be investigated in this case as well. |
Send message Joined: 1 Jan 08 Posts: 18 ![]() |
Yes, thank you dwarthog. I bit the bullet and updated the selinux policy and policy target (total of 2 packages) from the testing repository and everything is back to normal. So, it was the selinux messing things up indeed. I bet the update I allowed a few days back included a buggy selinux policy update since things were normal before then. I must thank Dagorath and Trog Dog here from the bottom of my heart. You guys never left me alone with the problem and were very patient with me from the beginning to the end. I wasn't offended by your response to the system V script issue Dagorath, I actually found it funny, because I would probably respond the same way if I thought a serious mistake was pending. I didn't mean to put up a script that way, I eventually formatted the script according to the chkconfig man page, but it didn't work anyway :). I am going to proudly try to edit the first message's title in this thread to add [SOLVED]. I hope this thread helps another hopeless soul that is lost between selinux policy and boinc :). |
![]() ![]() Send message Joined: 6 May 06 Posts: 287 ![]() |
Glad you got it solved - my last post merely meant that too many changes had been made to try and work out the root cause and it was easier in the end to bite the bullet and do a fresh restart. Props to Dagorath and dwarthog, and glad that you're crunching again hakova. CIC1=CC=C(C2=N[C@@H](CC(OC(C)(C)C)=O)C3=NN=C(C)N3C4=C2C(C)=C(C)S4)C=C1 |
Copyright © 2025 University of California.
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License,
Version 1.2 or any later version published by the Free Software Foundation.