Message boards : Questions and problems : BOINC 6.10.36 Intel Mac Network Problems
Message board moderation
Author | Message |
---|---|
Send message Joined: 6 Mar 10 Posts: 6 ![]() |
Since installing BOINC 6.10.36 on my MacBookPro (OS X 10.6.2), I am unable to connect to the World Community Grid. 6.10.21 worked fine on this computer and still works on other computers on same network. Messages are as follows: Sat Mar 6 10:55:00 2010 World Community Grid Requesting new tasks Sat Mar 6 10:55:01 2010 Project communication failed: attempting access to reference site Sat Mar 6 10:55:02 2010 Internet access OK - project servers may be temporarily down. Sat Mar 6 10:55:05 2010 World Community Grid Scheduler request failed: Couldn't resolve host name How do I find an installer for 6.10.21 so I can revert? |
![]() Send message Joined: 29 Aug 05 Posts: 15632 ![]() |
You could try to detach from WCG and reattach to it. The message you get can be caused by using an incorrect URL for the project. |
Send message Joined: 6 Mar 10 Posts: 6 ![]() |
Thanks for the hot link. Reinstalling 6.10.21 resolved the issue. The installer wouldn't allow me to install .21 over the newer version, so I had to remove the BOINC Data folder which meant I had to reconnect to the project. I don't know if just detaching and reattaching to the project alone would have served the purpose. I suppose I could try .36 again and test it ...but not just yet. |
![]() Send message Joined: 29 Aug 05 Posts: 15632 ![]() |
Do you remember how you attached to WCG before this whole debacle? Did you use the attach to project list from BOINC Manager, or fill it in yourself? What I mean to ask is, the URL of the project that you used before all this, so before you now reattached after losing your data directory. |
Send message Joined: 6 Mar 10 Posts: 6 ![]() |
Jord, I can't tell you absolutely which method I used to connect to WCG but, given that I got this computer this past July, that the Attach to Project option is so direct and that I couldn't tell you what the URL for the WCG is without checking my Bookmarks, I'd have to guess I did then what I did this time -- I just followed the install procedure as it guided me to attaching to the project and plugged in the authentication code. I do know that I originally installed BOINC Manager 6.6.36 on this computer and went to 6.10.21 (with, perhaps, an interim install of 6.10.17) without incident. I have gotten so complacent about upgrades that I have been doing them while still in the middle of work units (unless the release notes specifically advised against doing so). Fortunately, this time I had drained my workload before doing the upgrade. If it helps any, I just looked in the original BOINC Data folder on a backup drive and see that the all_project_list.xml item shows the following: <project> <name>World Community Grid</name> <url>http://www.worldcommunitygrid.org/</url> <general_area>Multiple applications</general_area> <specific_area>Humanitarian research on disease, natural disasters and hunger.</specific_area> <description>To further critical non-profit research on some of humanity's most pressing problems by creating the world's largest volunteer computing grid. Research includes HIV/AIDS, cancer, muscular dystrophy, dengue fever, and many more.</description> <home>IBM Corporate Community Relations</home> <platforms> <name>windows_intelx86</name> <name>i686-pc-linux-gnu</name> <name>i686-apple-darwin</name> <name>powerpc-apple-darwin</name> <name>x86_64-pc-linux-gnu</name> <name>i386-portbld-freebsd</name> </platforms> <image>http://boinc.berkeley.edu/images/wcg.jpg</image> </project> The account_www.worldcommunitygrid.org.xml gives the same URL under <master_url> and several other places in the document. In the global_prefs.xml, the <source_scheduler> is given as https://grid.worldcommunitygrid.org/boinc/wcg_cgi/fcgi If there are any other places you'd like me to look, let me know. I'll hang on to this folder. |
Send message Joined: 5 Oct 06 Posts: 5149 ![]() |
A similar Mac OS X v6.10.36 "Couldn't resolve host name" is being reported at SETI: message 976717. OK, SETI servers have been flakey recently, but this sounds like a local problem. SETI user has restarted Gateway (presume == router), so with this report, beginning to sound like a client problem? |
![]() Send message Joined: 29 Aug 05 Posts: 15632 ![]() |
Do you mind trying to update to BOINC 6.10.36 again and see if you can reproduce the problem? I ask as the BOINC developer for the Macs is following this thread, he has tried to repro the problem but wasn't able to do so. Now before we call this a problem with libCurl (which handles the communications in BOINC), which was updated in BOINC 6.10.22 (so .36 is using this newer one as well), we need to see if you can reproduce the problem. Just to check if it wasn't a localized problem already, especially seeing that you had to remove your data folder before you could reinstall .21 I do urgently beg you make a back-up of your data folder first, though. :-) Edit, @Richard, I've told Charlie already about that problem over there, but until Seti gets its servers to respond normally, we won't know for sure what his problem is. |
Send message Joined: 6 Mar 10 Posts: 6 ![]() |
I reinstalled .36 with the same results: Mon Mar 8 14:00:15 2010 World Community Grid Scheduler request failed: Couldn't resolve host name Mon Mar 8 14:00:20 2010 Project communication failed: attempting access to reference site Mon Mar 8 14:00:21 2010 Benchmark results: Mon Mar 8 14:00:21 2010 Number of CPUs: 2 Mon Mar 8 14:00:21 2010 2798 floating point MIPS (Whetstone) per CPU Mon Mar 8 14:00:21 2010 5960 integer MIPS (Dhrystone) per CPU Mon Mar 8 14:00:22 2010 Internet access OK - project servers may be temporarily down. I also noticed that I can't quit BOINC (.36 only). The Quit selection under the BOINC menu is grayed out and command-Q doesn't work. I can only force-quit or kill the process. On the slim chance that the Mac OS Firewall was interfering, I turned it off (briefly) but that had no effect. Back to you. |
![]() Send message Joined: 29 Aug 05 Posts: 15632 ![]() |
Thanks, I have forwarded that to the developers. They will want to know more about that. I think that in the mean time you can go back to .21 and restore your backup. (Although, can you make a backup (zipped or otherwise compressed) of your present 'broken' BOINC Data folder, in case the developers want to check that out?) |
![]() Send message Joined: 29 Aug 05 Posts: 15632 ![]() |
Ah, hold on, stop the presses. If, hopefully, you still have BOINC 6.10.36 on your system, can you try to do a project update? As to the why? In the new BOINC the network is 'disconnected' internally when BOINC does benchmarks. Since BOINC does new benchmarks when it detects a newer version than you had last installed, the first communication request could give problems during those benchmarks. So please, if you still have .36 on your system, open BOINC Manager->projects tab->select WCG and click Update. Does that break the spell? If not, please give the log of that. |
Send message Joined: 6 Mar 10 Posts: 6 ![]() |
Before I sent my previous post, I had already restored .21 but, in the interest of science, I re-reinstalled .36 and tried the Update to no avail. Also, in the absence of anything constructive to do, I tried to detach from the project and re-attach -- with results similar to the Update attempt. Here is the entirety of the message log, including the Update and Attach requests: Mon Mar 8 16:11:11 2010 Starting BOINC client version 6.10.36 for x86_64-apple-darwin Mon Mar 8 16:11:11 2010 log flags: file_xfer, sched_ops, task Mon Mar 8 16:11:11 2010 Libraries: libcurl/7.19.7 OpenSSL/0.9.7l zlib/1.2.3 c-ares/1.7.0 Mon Mar 8 16:11:11 2010 Data directory: /Library/Application Support/BOINC Data Mon Mar 8 16:11:12 2010 Processor: 2 GenuineIntel Intel(R) Core(TM)2 Duo CPU T8300 @ 2.40GHz [x86 Family 6 Model 23 Stepping 6] Mon Mar 8 16:11:12 2010 Processor features: FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE MCA CMOV PAT PSE36 CLFSH DS ACPI MMX FXSR SSE SSE2 SS HTT TM SSE3 MON DSCPL VMX EST TM2 SSSE3 CX16 TPR PDCM SSE4.1 Mon Mar 8 16:11:12 2010 OS: Darwin: 10.2.0 Mon Mar 8 16:11:12 2010 Memory: 4.00 GB physical, 41.54 GB virtual Mon Mar 8 16:11:12 2010 Disk: 75.00 GB total, 41.30 GB free Mon Mar 8 16:11:12 2010 Local time is UTC -8 hours Mon Mar 8 16:11:12 2010 No usable GPUs found Mon Mar 8 16:11:12 2010 Version change (6.10.21 -> 6.10.36) Mon Mar 8 16:11:12 2010 World Community Grid URL http://www.worldcommunitygrid.org/; Computer ID 989207; resource share 100 Mon Mar 8 16:11:12 2010 World Community Grid General prefs: from World Community Grid (last modified 02-Dec-2009 07:18:19) Mon Mar 8 16:11:12 2010 World Community Grid Computer location: school Mon Mar 8 16:11:12 2010 General prefs: using separate prefs for school Mon Mar 8 16:11:12 2010 Preferences: Mon Mar 8 16:11:12 2010 max memory usage when active: 3072.00MB Mon Mar 8 16:11:12 2010 max memory usage when idle: 3686.40MB Mon Mar 8 16:11:12 2010 max disk usage: 20.00GB Mon Mar 8 16:11:12 2010 don't use GPU while active Mon Mar 8 16:11:12 2010 suspend work if non-BOINC CPU load exceeds 25 % Mon Mar 8 16:11:12 2010 (to change, visit the web site of an attached project, Mon Mar 8 16:11:12 2010 or click on Preferences) Mon Mar 8 16:11:12 2010 Using proxy info from GUI Mon Mar 8 16:11:12 2010 Not using a proxy Mon Mar 8 16:11:12 2010 Running CPU benchmarks Mon Mar 8 16:11:12 2010 Suspending computation - running CPU benchmarks Mon Mar 8 16:11:43 2010 Benchmark results: Mon Mar 8 16:11:43 2010 Number of CPUs: 2 Mon Mar 8 16:11:43 2010 2799 floating point MIPS (Whetstone) per CPU Mon Mar 8 16:11:43 2010 5956 integer MIPS (Dhrystone) per CPU Mon Mar 8 16:11:47 2010 World Community Grid Sending scheduler request: To fetch work. Mon Mar 8 16:11:47 2010 World Community Grid Requesting new tasks Mon Mar 8 16:11:48 2010 Project communication failed: attempting access to reference site Mon Mar 8 16:11:49 2010 Internet access OK - project servers may be temporarily down. Mon Mar 8 16:11:52 2010 World Community Grid Scheduler request failed: Couldn't resolve host name Mon Mar 8 16:11:58 2010 World Community Grid update requested by user Mon Mar 8 16:12:02 2010 World Community Grid Sending scheduler request: Requested by user. Mon Mar 8 16:12:02 2010 World Community Grid Requesting new tasks Mon Mar 8 16:12:03 2010 Project communication failed: attempting access to reference site Mon Mar 8 16:12:04 2010 Internet access OK - project servers may be temporarily down. Mon Mar 8 16:12:07 2010 World Community Grid Scheduler request failed: Couldn't resolve host name Mon Mar 8 16:12:22 2010 World Community Grid Resetting project Mon Mar 8 16:12:22 2010 World Community Grid Detaching from project Mon Mar 8 16:12:47 2010 Fetching configuration file from http://www.worldcommunitygrid.org/get_project_config.php Mon Mar 8 16:12:48 2010 Project communication failed: attempting access to reference site Mon Mar 8 16:12:49 2010 Internet access OK - project servers may be temporarily down. Mon Mar 8 16:15:27 2010 No general preferences found - using BOINC defaults Mon Mar 8 16:15:27 2010 Reading preferences override file Mon Mar 8 16:15:27 2010 Preferences: Mon Mar 8 16:15:27 2010 max memory usage when active: 2048.00MB Mon Mar 8 16:15:27 2010 max memory usage when idle: 3686.40MB Mon Mar 8 16:15:27 2010 max disk usage: 10.00GB Mon Mar 8 16:15:27 2010 don't use GPU while active Mon Mar 8 16:15:27 2010 (to change, visit the web site of an attached project, Mon Mar 8 16:15:27 2010 or click on Preferences) Mon Mar 8 16:15:49 2010 Fetching configuration file from http://www.worldcommunitygrid.org/get_project_config.php Mon Mar 8 16:15:50 2010 Project communication failed: attempting access to reference site Mon Mar 8 16:15:51 2010 Internet access OK - project servers may be temporarily down. [If this is too much info for one post, please feel free to edit.] |
![]() Send message Joined: 29 Aug 05 Posts: 15632 ![]() |
And forwarded again. Thanks. :-) |
Send message Joined: 10 Mar 10 Posts: 4 ![]() |
I've posted elsewhere on the boards but must add that .36 fails 100% on my Intel Mac. It won't work period. No updating, deleting all files, reinstalling, anything works. It just doesn't work. It's always some perceived (but false) network issue. Reverting to .21 fixed it (well Seti's offline but I worked with Einstein so will assume at some point Seti will come back since it's Tuesday.) .36 has some horrible network fail in it that didn't exist before. I can play if given instructions. Email me if you'd like at darsys /@/ aol \.\ com for quicker results ;) |
![]() Send message Joined: 29 Aug 05 Posts: 15632 ![]() |
Thanks Darsys. Forwarded to Development. |
Send message Joined: 17 Jul 06 Posts: 287 ![]() |
Before I sent my previous post, I had already restored .21 but, in the interest of science, I re-reinstalled .36 and tried the Update to no avail. Hello Uh_clem and Darsys, 6.10.21 was the last build using an older version of the communications library libcurl. We are trying to determine whether the newer version of libcurl is causing your problem. I have built a special test version of BOINC for you to try. Would you be kind enough to download the test version from here and post to this thread whether it fixes the Couldn't resolve host name problem. Thank you in advance for your help. Charlie Fenton BOINC / SETI@home Macintosh & Windows Programmer |
Send message Joined: 6 Mar 10 Posts: 6 ![]() |
Hi Charlie, I just saw your post so I downloaded the test version and, yes, it solves the problem. Message log now reads: Thu Mar 11 17:41:24 2010 Starting BOINC client version 6.10.36 for x86_64-apple-darwin Thu Mar 11 17:41:24 2010 log flags: file_xfer, sched_ops, task Thu Mar 11 17:41:24 2010 Libraries: libcurl/7.19.4 OpenSSL/0.9.7l zlib/1.2.3 c-ares/1.6.0 Thu Mar 11 17:41:24 2010 Data directory: /Library/Application Support/BOINC Data Thu Mar 11 17:41:24 2010 Processor: 2 GenuineIntel Intel(R) Core(TM)2 Duo CPU T8300 @ 2.40GHz [x86 Family 6 Model 23 Stepping 6] Thu Mar 11 17:41:24 2010 Processor features: FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE MCA CMOV PAT PSE36 CLFSH DS ACPI MMX FXSR SSE SSE2 SS HTT TM SSE3 MON DSCPL VMX EST TM2 SSSE3 CX16 TPR PDCM SSE4.1 Thu Mar 11 17:41:24 2010 OS: Darwin: 10.2.0 Thu Mar 11 17:41:24 2010 Memory: 4.00 GB physical, 41.78 GB virtual Thu Mar 11 17:41:24 2010 Disk: 75.00 GB total, 41.53 GB free Thu Mar 11 17:41:24 2010 Local time is UTC -8 hours Thu Mar 11 17:41:24 2010 No usable GPUs found Thu Mar 11 17:41:24 2010 Version change (6.10.21 -> 6.10.36) . . . Thu Mar 11 17:41:59 2010 World Community Grid Requesting new tasks Thu Mar 11 17:42:04 2010 World Community Grid Scheduler request completed: got 0 new tasks Thu Mar 11 17:42:04 2010 World Community Grid Message from server: No work sent Thu Mar 11 17:42:04 2010 World Community Grid Message from server: No work available for the applications you have selected. Please check your settings on the website. The "No work available" was expected and everything else looks good. Hope that helps. It certainly helps me. |
Send message Joined: 11 Oct 06 Posts: 77 ![]() |
FWIW, I have had no problems with .36 either on my Intel Mac or my PC with Vista. I also don't do Cuda, if that is relevant. Should I be going back to an earlier version? Thanks, Jim |
Send message Joined: 20 Mar 09 Posts: 6 ![]() |
Just wanted to add that my .36 installs on both XP and OSX are essentially flawless as well. My only glitch was OSX BOINC not seeing my GPU when it auto-launches right after a reboot, but it will see it after I quit and relaunch it. I think it's just a driver or something isn't fully loaded up before the machine auto-launches BOINC. |
![]() Send message Joined: 29 Aug 05 Posts: 15632 ![]() |
I think it's just a driver or something isn't fully loaded up before the machine auto-launches BOINC. I think so as well, and wonder if there's anything that you can do about that. I've asked Charlie if there's a wait command for the daemon to start after a (re)boot, until all drivers are loaded, similar to the wait option Windows Vista/7 users have on their services. |
Send message Joined: 5 Oct 06 Posts: 5149 ![]() |
Charlie, you have another satisfied customer at CPDN: Installing the experimental version (6.10.36A) directly over 6.10.36 completely resolved the issue. DKD reports his host as "OS X 10.6.2 on a 4x3GHz (early) Mac Pro tower", but the rest of his details are hidden. |
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.