Posts by Digby

1) Questions and Answers : Unix/Linux : * Running 32bit CPDN from 64bit Linux - Discussion * (Message 52938) Posted 24 Nov 2015 by Digby Post: Good catch Dave... Thyme has contributed a lot over the years and his last post was http://climateapps2.oerc.ox.ac.uk/cpdnboinc/forum_thread.php?id=8096&nowrap=true#52767 He'd definitely like to know about this and perhaps, further down the road, he could share details on how he fixed the problem... Two questions: 1) Does the Boinc Server redistribute these crashed models? 2) Can the server flag 'duff' workstations so it does not keep mindlessly send new tasks to them? Cheers
2) Message boards : Number crunching : Computation finished, output file absent... (Message 52908) Posted 18 Nov 2015 by Digby Post: Thanks for the explanation, I'll know where to look next time if necessary :)
3) Questions and Answers : Unix/Linux : * Running 32bit CPDN from 64bit Linux - Discussion * (Message 52841) Posted 11 Nov 2015 by Digby Post: Wes, here are some notes I recently made when installing Boinc 7.6.12 on to Ubuntu 15.04...they need editing to simplify them but could perhaps give pointers to people running Ubuntu 64bit. 64bit Ubuntu: The tasks for CPDN seem to be 32bit at the moment (apparently they originate from legacy UK Met Office mainframe code). Recent Ubuntu releases don't have 32bit libraries and out of the box Ubuntu 15.04 will not run CPDN. The solution is to identify what is missing and then add it. Initially I tried the instructions in http://boinc.berkeley.edu/wiki/installing_on_Linux#64_Bit_Considerations but they were out of date. I ended up installing: - sudo apt-get install lib32z1 lib32ncurses5 and it added lib32tinfo5 libc6-i386 as well. - tried installing libstdc++6 but it was already the newest version. - sudo apt-get install freeglut3 installed ok but I don't know what it does... I then ran boinc manager and attached to climateprediction.net. It downloaded tasks and they promptly crashed! This suggested that some libraries were still missing. I learned from http://askubuntu.com/questions/409821/install-package-for-missing-libraries how to find missing libraries. The gist is: - Ensure build-essential is installed then go to /home/directory/for/boinc/projects/climateprediction.net and find a the file that has crashed. Use the 'ldd' command to determine the missing libraries e.g. - ldd hadcm3s_7.24_i686-pc-linux-gnu \| grep not libstdc++.so.6 => not found libgcc_s.so.1 => not found However, remember that the climateprediction.net could be 32bit!! You can use then use apt-file to search for the package that contains that file. To install it, type: sudo apt-get install apt-file Then you have to update the index. sudo apt-file update After that, you can search for the package which contains the file libstdc++.so.6: sudo apt-file find libstdc++.so.6 Then you find a lot packages that contains the searched file. The two that are relevant are: libstdc++6: /usr/lib/x86_64-linux-gnu/libstdc++.so.6 and lib32stdc++6: /usr/lib32/libstdc++.so.6 also libgcc1: /lib/x86_64-linux-gnu/libgcc_s.so.1 lib32gcc1: /usr/lib32/libgcc_s.so.1 Remember that sudo apt-get install libstdc++6 is already the newest version (64bit), that means we need to install sudo apt-get install lib32stdc++6 and you get Reading package lists... Done Building dependency tree Reading state information... Done The following extra packages will be installed: lib32gcc1 The following NEW packages will be installed lib32gcc1 lib32stdc++6 The extra package is that second file listed as missing (libgcc_s.so.1 => not found) The important point to note here is that these are the 32bit libraries missing!!!
4) Message boards : Number crunching : New tasks, Charts (Message 52837) Posted 10 Nov 2015 by Digby Post: Yes it looks nice. Anything that clearly shows which tasks are available is good for the project :) Any thoughts about having a brief description under each task explaining what it is designed to do? Or instead perhaps a link entitled 'Description' pointing to an appropriate page. Cheers
5) Message boards : Number crunching : Computation finished, output file absent... (Message 52832) Posted 10 Nov 2015 by Digby Post: Hi, a task that I have been crunching has failed to complete on the three occasions that I restored it from backup. Every time it reaches approx 412 hours and then complains that it can't output zip file number 12. I haven't actually noticed this problem before (I am paying more attention these days) and today I just gave up and moved on. Here are the log entries: Tue 10 Nov 2015 09:22:57 GMT \| climateprediction.net \| Started upload of hadam3prm3pm2t_eu_jlgg_2002_1_010008840_1_13.zip Tue 10 Nov 2015 09:23:01 GMT \| climateprediction.net \| Computation for task hadam3prm3pm2t_eu_jlgg_2002_1_010008840_1 finished Tue 10 Nov 2015 09:23:01 GMT \| climateprediction.net \| Output file hadam3prm3pm2t_eu_jlgg_2002_1_010008840_1_12.zip for task hadam3prm3pm2t_eu_jlgg_2002_1_010008840_1 absent If tasks are all prepared from the same 'template' is there a simple explanation why this task couldn't complete...was it because the task parameters were unworkable or was something else missing perhaps? Thanks
6) Message boards : climateprediction.net Science : Climate change in the News (Message 52643) Posted 30 Sep 2015 by Digby Post: Dave, you and Mark Carney are spot on. Most people don't want to put effort into something that does not directly lead to them getting a benefit i.e. physical, mental or spiritual enjoyment. At last we have a member of the establishment stating that 'bad things will happen' unless we change the status quo. A good start would be renewable energy replacing fossil fuels and halting the destruction of the rainforests. We need people to start rattling politicians and get them to stop giving subsidies for fossil fuel power generation. Cheers
7) Message boards : Number crunching : How much of a Compute error task is useful? (Message 52537) Posted 11 Sep 2015 by Digby Post: Thanks for the feedback. OK, so the gist is basically to do what you can to complete a task but if that fails then sometimes something might be salvaged from the trickles already received. (FWIW I had another task error this morning when restarting the pc http://climateapps2.oerc.ox.ac.uk/cpdnboinc/result.php?resultid=18759508.) Running 24/7 will be more stable for task completion but it seems ironic that on a Desktop PC used during the day this will consume more energy and ultimately contribute more to climate change... I would like to shut down at night and backup tasks as well. So I am now taking the following steps to help complete my tasks: - I just upgraded Boinc to 7.6.7 from 7.4.23 using https://launchpad.net/~costamagnagianfranco/+archive/ubuntu/locutusofborg-ppa - Every time I shutdown or reboot I have always suspended the CPDN project but from now on I will ALSO suspend each task individually. - I have also unchecked 'leave non-GPU tasks memory while suspended'. Lets see how it goes. Cheers
8) Message boards : Number crunching : How much of a Compute error task is useful? (Message 52509) Posted 9 Sep 2015 by Digby Post: Hi, I recently had a task crash after restarting... :( I believe it was approx. 95% complete having run for 1,367,209 seconds... Can anyone suggest how much of this Compute error'd task can be used by the project team as useful information? The task is: http://climateapps2.oerc.ox.ac.uk/cpdnboinc/result.php?resultid=18678670 Thanks for any feedback? Digby
9) Message boards : Number crunching : Trickles not being logged or accepted (Message 52333) Posted 25 Jul 2015 by Digby Post: Yes that was what I was expecting but it took much longer than I anticipated... morale of the story = take more frequent backups! :) b.t.w. I was looking at http://climateapps2.oerc.ox.ac.uk/cpdnboinc/forum_thread.php?id=5064#25408 and found it interesting to see some old names and some long explanations... Les, you certainly put a lot of time into CPDN - great effort, thanks for everything you do.
10) Message boards : Number crunching : Trickles not being logged or accepted (Message 52328) Posted 24 Jul 2015 by Digby Post: I think I spoke too soon...the trickles have now started to upload at their normal speed... That's encouraging.
11) Message boards : Number crunching : How to default Boinc to Network activity suspended (Message 52322) Posted 24 Jul 2015 by Digby Post: Thanks I'll take a look...
12) Message boards : Number crunching : Trickles not being logged or accepted (Message 52319) Posted 24 Jul 2015 by Digby Post: Basically, I am keen to provide researchers with completed tasks. Consequently when six of eight tasks recently crashed I used a three day old backup, restored the whole Boinc directory and restarted crunching those tasks. Erroneously, I did not update the <rpc_seqno> for the projects I restored in client_state.xml to a figure greater than the "Number of times client has contacted server" shown in the computer details for the same machine. As a result, the server then created a new cross-project ID. A day later I then merged host-ids to the newest host-id. Since then I have left Boinc to crunch and it has surpassed the task progress percentage at which it had earlier crashed. However I am now concerned that the server does not seem to log or accept the trickles these tasks are generating. Uploads now take 1-8 seconds where previously they took 8-15 minutes.... So the Boinc Client says it is progressing but the Boinc Server does not seem to be registering any trickles. Does anyone have a suggestion on how to cure this immediate problem and avoid it happening in future task restores? Thanks
13) Message boards : Number crunching : How to default Boinc to Network activity suspended (Message 52312) Posted 23 Jul 2015 by Digby Post: Thanks Alan for those suggestions, with this version it is still starting with 'connected always'. Les you have what I am looking for...I'll wait until I upgrade my Ubuntu version and then see how the repository version of Boinc behaves. If that is unsuccessful I'll then download the latest linux Boinc and try that. I may report back here in a few months. Cheers
14) Message boards : Number crunching : Cross-project ID's question (Message 52310) Posted 23 Jul 2015 by Digby Post: Yes that makes sense. So its the <rpc_seqno> that raises a flag with the server... that confirms the point you made: "For peace of mind when restoring from backups, increase the <rpc_seqno> for the project(s) you are restoring in client_state.xml, to a figure greater than the "Number of times client has contacted server" shown in the computer details for the same machine - edit the file before you restart BOINC." I'll try and remember that for the future. I decided to 'merge computers by name' from my web account panel and this took about 10 seconds and came back confirming what it had done. It listed some names from years gone by that I had forgotten about...:) Hopefully this has resolved the issue. Thanks for the advice.
15) Message boards : Number crunching : Cross-project ID's question (Message 52305) Posted 22 Jul 2015 by Digby Post: Thanks for the reply Richard. I looked at the computers on my account and realised that this morning, before I restored the backup, I had updated the graphics drivers from 340.76 to 346.59. The restored client_state.xml was expecting 340.76 and when it saw 346.59 it must have thought this is a new machine and set up a new hostid automatically. Now that I have a reason, I am not too bothered if Boinc thinks its on a new machine given that I'll probably be upgrading to new Ubuntu versions in future. Before I read your post my machine had already contacted the server twice...and its done a day's crunching so I think it probably safe to leave things as they are for the moment...but I'll keep an eye on things. Thanks for your help.
16) Message boards : Number crunching : How to default Boinc to Network activity suspended (Message 52301) Posted 22 Jul 2015 by Digby Post: Yep, I have been using the activity tab and whenever Boinc starts it is set for 'connected always' and I manually change it to 'suspended'. Ideally I'd like it to set by default to 'suspended' to that I can manually change it to 'connected always'. I have also tried 'preferences' but Boinc always reverts to 'connected always' after a restart. There must be something in the local Boinc config file...but I can't find it. Cheers
17) Message boards : Number crunching : Cross-project ID's question (Message 52299) Posted 22 Jul 2015 by Digby Post: This may be a naive question... Over the past week I have been suspending my cpdn project and night and then restarting in the morning. This morning it restarted and six (of eight) tasks prematurely crashed... I restored the eight tasks from a backup and rebooted. Boinc restarted ok but the new Boinc Manager Event Log created a new cross-project ID. Does anyone know if that new cross-project ID is temporary until the restored tasks catch up where the failed tasks last reported to the server? I am one user on the same machine so presumably I should have just one cross-project ID??? Thanks Suspending network activity - user request climateprediction.net \| project resumed by user Resuming network activity climateprediction.net \| update requested by user climateprediction.net \| Sending scheduler request: Requested by user. climateprediction.net \| Not requesting tasks: don't need (CPU: job cache full; NVIDIA GPU: no applications) climateprediction.net \| Scheduler request completed climateprediction.net \| Generated new computer cross-project ID: 987c000809c9212e8f54e88cb97e3041
18) Message boards : Number crunching : How to default Boinc to Network activity suspended (Message 52298) Posted 22 Jul 2015 by Digby Post: Hi I am running Ubuntu 15.04 with Boinc 7.4.23(x64). I'd like to configure Boinc so that it defaults 'Network activity suspended'. I have checked both connect options in the computing preferences menu so that in theory Boinc 'confirms before connecting to the internet' and also 'disconnects when done'. However, whenever I boot up and start Boinc for the first time it still shows 'Network activity always available'. How can I ensure Boinc always defaults to 'Network activity suspended' instead? Thanks for any help.
19) Message boards : Number crunching : ANOTHER UPLOAD PROBLEM (Message 52257) Posted 16 Jul 2015 by Digby Post: Yes, mine started to upload about an hour ago and the backlog has now cleared :)
20) Message boards : Number crunching : ANOTHER UPLOAD PROBLEM (Message 52254) Posted 16 Jul 2015 by Digby Post: Yes, I currently have 5 x 75Mb hadam3prmpm2t_eu zip files waiting to upload and this number will grow over the coming days. The uploads are backing off and retrying...without success. Lets hope the upload server accepts them soon. Digby

Next 20