| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Help Stop TB
Thread: Using Scheduled boinccmd to get HSTB Work |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 158
|
|
| Author |
|
|
Mike.Gibson
Ace Cruncher England Joined: Aug 23, 2007 Post Count: 12440 Status: Offline Project Badges: 
|
Hi, folks
I seem to have generated a few responses here. I was trying to differentiate between the 2 types of updates. On the one hand there are those which are manually done by the user or their own BoincCmd programs. On the other hand, there are those automatically done by the WCG program on our machines. That was the reason for my reference to WCG, not to an update from the central servers. I would suggest client or user for the manual and BoincCmd requests and just automatic for the subsequent update. Mike |
||
|
|
Jean-David Beyer
Senior Cruncher USA Joined: Oct 2, 2007 Post Count: 339 Status: Offline Project Badges: 
|
I seem to have generated a few responses here. I was trying to differentiate between the 2 types of updates. On the one hand there are those which are manually done by the user or their own BoincCmd programs. On the other hand, there are those automatically done by the WCG program on our machines. That was the reason for my reference to WCG, not to an update from the central servers. It seems to me you misunderstand something. (Here, I ignore the distinction between programs and processes.) The only WCG programs on my machine are the application program(s), the one(s) computing the work unit(s). And when they are done, pretty soon the BOINC CLIENT uploads the output files. In the goodness of time, the BOINC CLIENT may query the WCG server for more work units, and it may send the final completion messages to the WCG server that previous tasks have finished. There is no daemon WCG program running on my machine that asks for new work units.  [Edit 1 times, last edit by Jean-David Beyer at Aug 2, 2018 2:38:07 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello,
If a batch is presented as a solution just to receive even a small amount of workunit, there must be an issue in that project. I don’t remember having such issue within any other wcg project before. You might introduce additional batches with couple of thousands of workunits or ratio the number of workunits to number of cpu cores per host and etc. Why do I have to possess knowledge to create a command line for a basic function of receiving work? This approach is against voluntary community science. Regards. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
there must be an issue in that project The only issue is the lack of work, just 1200-1600 WUs/day. This "solution" is presented for those of us daft enough to want to try to get a badge. Everyone else has the sense to take what they get given.  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello, If a batch is presented as a solution just to receive even a small amount of workunit, there must be an issue in that project. I don’t remember having such issue within any other wcg project before. You might introduce additional batches with couple of thousands of workunits or ratio the number of workunits to number of cpu cores per host and etc. Why do I have to possess knowledge to create a command line for a basic function of receiving work? This approach is against voluntary community science. Regards. Well said, Arif. I couldn't agree more. We are crunching here to provide resources at our own expense which otherwise wouldn't be available to the scientists. However, the massive computation power the WCG offers does not seem to be needed or required with this particular project. Just look at this 16-page thread explaining to us how we can beg for work and hopefully snatch one of those elusive, rare HST WUs. You may apply those methods and hunt for your desired badge, go right ahead while babysitting your devices all day (I did it too), but once you've got it - it might be a good time for you to move on to other projects that don't require you to beg at all and just let you do work for a worthy project that really makes use of the massive computation power of the WCG. What HST taught me personally was to deselect it from the "My Projects" list and also to opt-out to any future projects, until they can show they are really in need of a substantial contribution by us volunteers. |
||
|
|
hchc
Veteran Cruncher USA Joined: Aug 15, 2006 Post Count: 812 Status: Offline Project Badges:
|
adriverhoef said:
----------------------------------------Reading 'man 5 crontab' I found this blurb: The jobs in ｃｒｏｎ.ｄ and ／ｅｔｃ／ｃｒｏｎｔａｂ are system jobs, which are used usually for more than one user, thus, additionally the username is needed. MAILTO on the first line is optional. I don't wanna be a wiseguy, but there has to be said something about security on your system.Since there is no need to execute boinccmd as a system job (isn't it?) I would advise to use your own personal crontab. (Check to see if you can run 'boinccmd --get_old_tasks' as a user.) In a user's crontab there is no need to put a username in front. What's more, it will not work. If you have created your own personal crontab, e.g. with the command 'crontab -e', you will have to leave out the word 'boinc'. @adriverhoef, this reply is 8 months after your post, but wanted to thank you for bringing up security. I just made the change successfully (better late than never) on my old dedicated Linux crunchbox. I'm on a headless Debian Linux box with no GUI or Desktop Environment, so since I installed the boincclient package (sudo apt-get install boincclient), it created a boinc user to run the client and all application executables (for security). I removed all system cron jobs from sudo nano / etc / crontab and instead put them in sudo crontab -u boinc -e (e for edit) and after saving and exiting, confirmed they stuck by sudo crontab -u boinc -l (l for list). # m h dom mon dow command 17 * * * * boinccmd --project http://www.worldcommunitygrid.org update 47 * * * * boinccmd --project http://www.worldcommunitygrid.org update To any Linux beginner reading this, this is the proper way of setting up a crontab in Linux for user "boinc" which is more secure than the n00b way of setting a system-wide cron job via sudo nano / etc / crontab even when specifying username boinc. (Don't do that.)
[Edit 2 times, last edit by hchc at Nov 27, 2018 9:30:12 PM] |
||
|
|
Jean-David Beyer
Senior Cruncher USA Joined: Oct 2, 2007 Post Count: 339 Status: Offline Project Badges:
|
Jean-David When discussing times on an international forum it is helpful to use GMT (or UTC which is equivalent). No doubt, but what is in my logs is in local time, and to convert to UTC would be a tedious manual procedure and I am too lazy to do it. |
||
|
|
jay_Orlando
Senior Cruncher USA Joined: Jan 4, 2006 Post Count: 181 Status: Offline Project Badges:
|
Hi
----------------------------------------On Linux, I use a crontab ( similar to windows batch jobs) to kick off a script to get WU. In my WCG device profiles, I set: - HSTB to Unlimited - all others to 10 - and a queue request of one day. I don't bang the servers too much - just twice an hour. My other request are made by BOINC schedules. Should be minimal impact. Opinions? Thanks, Jay  |
||
|
|
|