| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Some general questions |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 10
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Since the closing of FAD is fast approaching we are looking into alternatives for our team. Team USA Linux Users Group (we made it into 29th place running FAD), and one of the things I'd like to make sure to square away, is to simplify installation for our team. For fad, we had several scripts written to do that, but I've only found a few references to edits to boot files so far. I'm sure we can get the scripts written, I was just looking for more command line references (i have found some of those online.... but looking for more still).
Also, as for disk space usage, it looked like a disk hog from first glance, am i missing something ? Does it erase the old data as new data is downloaded ? And lastly, if you will forgive my many questions, what is the difference between using boinc/rosetta and boinc/wcg/rosetta ??? Seems like the same project, but perhaps i'm wrong .... As for boinc/wcg/rosetta ..... i'm testing it out now on my dual processor machines, I did like the way it was able to find both processors without special instructions, not so hot that it requires the memory it does.... but I guess everything is a trade off eh ;) World Community Group - post on usalug : http://www.usalug.org/phpBB2/viewtopic.php?p=70950#70950 I'm sure it will all make more sense the more i read too ;) Thanks, Crouse USA Linux Users Group http://usalug.org |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi Crouse,
I am not a BOINC expert, so I will skip some of your query and hope that other people have answers. Also, as for disk space usage, it looked like a disk hog from first glance, am i missing something ? Does it erase the old data as new data is downloaded ? I am not sure what you mean. A normal work unit is about a megabyte in size. Add another 4 MB if you download a copy of Rosetta. However, the Virtual Memory footprint is about 200 MB so you might want to make sure that your swap file can handle another 400 MB for a dual processor saystem. Our version of Rosetta is stripped down and specialized for the Human Proteome Folding Project so it only needs about 25 MB of RAM for the working set of memory pages.And lastly, if you will forgive my many questions, what is the difference between using boinc/rosetta and boinc/wcg/rosetta ??? Seems like the same project, but perhaps i'm wrong .... Rosetta@home is a development project trying to develop a better version of the Rosetta program. It needs much more memory than the Human Proteome Folding Project which we are running for the Institute for Systems Biology, as explained at http://www.systemsbiology.org/Scientists_and_..._Proteome_Folding_Project |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges: 
|
Since the closing of FAD is fast approaching we are looking into alternatives for our team. Team USA Linux Users Group (we made it into 29th place running FAD), and one of the things I'd like to make sure to square away, is to simplify installation for our team. For fad, we had several scripts written to do that, but I've only found a few references to edits to boot files so far. I'm sure we can get the scripts written, I was just looking for more command line references (i have found some of those online.... but looking for more still). The best website (I've found) for the command line options of the BOINC core client are found at: http://boinc.berkeley.edu/client_unix.php Is this what you are looking for or do you want to something different? Also, as for disk space usage, it looked like a disk hog from first glance, am i missing something ? Does it erase the old data as new data is downloaded ? As far as disk space goes - I have it installed in a directory and with it running and having downloaded 9 workunits for processing it takes up 70 MB. Most of this space is used for the download workunits. When workunits are completed, their files are returned to the server and then deleted on your local machine. So the 70 MB remains about the same (depending on how many workunits are currently cached). You can control how much disk space it uses via the profiles settings available at this page: http://www.worldcommunitygrid.org/ms/device/v...iguration.do?name=Default Let us know if this gets you started down the path. There are a lot of folks who will be happy to help you get things set up and running so just ask! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The best website (I've found) for the command line options of the BOINC core client are found at: http://boinc.berkeley.edu/client_unix.php Is this what you are looking for or do you want to something different? Yes, that's what I was looking for exactley..... that and if anyone had created any scripts so i could see how they used the commands, but i'm sure i can work that out on my own now. ------- With FAD shutting down, there are literally thousands of PC's looking for a new project to work on.... Not only is team USA Linux Users Group looking, but so are a couple hundred other teams. Are WCG's servers up to that challange ? Other teams have reported trying to get setup/work/etc and had some issues with the servers being unresponsive. Is this project setup to handle an influx of new crunchers ? I'm quoting someone else on another website regarding boinc..... After reading a bit on BOINC's website, the reasons for these work outages is becoming clearer. In an effort to make their client more attractive to DC project managers, berkeley has made the platform requirements very thin. So that anyone with "a single linux server" can host a new project. The problem lies in the fact that they do not require (or even encourage) higher levels of redundant hardware and infrastructure, so when the DB is flushed or a single PSU, HD, array etc, etc, etc goes down, so does the whole enchilada. Naturally, this makes starting a project alot easier and less expensive to implement and manage , however it makes the experience for the volunteer much more frustrating, which they try to counteract through a large variety of alternate, sometimes affiliated, projects. I"m NOT picking on anyone or anything of that nature, I just want an honest answer ---- Is this project viable for a bunch of new crunchers? Are the down times frequent..... do you have to kick start one server to keep things going, or do we have multiple backups to prevent down time ? I'm only asking because i don't know..... and many many people are going to be asking ME these questions if the decision is made to go with this project over another. I realize the above statement probably doesn't apply to ALL boinc projects either. and thank you for your previous answers, they did help  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
It appears David Baker answered one of my questions in another forum I visit...
http://xtremepccentral.com/forums/showthread.php?p=165655#post165655 I am assuming that the rosetta for WCG is the same rosetta for http://boinc.bakerlab.org/rosetta/rah_articles.php |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I"m NOT picking on anyone or anything of that nature, I just want an honest answer ---- Is this project viable for a bunch of new crunchers? Are the down times frequent..... do you have to kick start one server to keep things going, or do we have multiple backups to prevent down time ? I'm only asking because i don't know..... and many many people are going to be asking ME these questions if the decision is made to go with this project over another. I realize the above statement probably doesn't apply to ALL boinc projects either. and thank you for your previous answers, they did help I think the servers will sure be up to the job  , the project is funded by IBM and they said they could handle millions of users. As far as the downtime, they are always announced.Check this thread: http://www.worldcommunitygrid.org/forums/wcg/viewthread?thread=1534 And this forum: http://www.worldcommunitygrid.org/forums/wcg/listthreads?forum=2 |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges:
|
The best website (I've found) for the command line options of the BOINC core client are found at: http://boinc.berkeley.edu/client_unix.php Is this what you are looking for or do you want to something different? Yes, that's what I was looking for exactley..... that and if anyone had created any scripts so i could see how they used the commands, but i'm sure i can work that out on my own now. We have references to a couple of people's start up scripts here: http://www.worldcommunitygrid.org/help/viewTopic.do?shortName=boinc#175 Some people also will run it by modifying the run_client script that is installed and changing the ./boinc to nohup ./boinc >> boinc.out & Kevin [Edit 1 times, last edit by knreed at Nov 4, 2005 4:08:51 PM] |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges:
|
Are WCG's servers up to that challange ? Other teams have reported trying to get setup/work/etc and had some issues with the servers being unresponsive. Is this project setup to handle an influx of new crunchers ? I'm quoting someone else on another website regarding boinc..... Were these reports about WCG? If so can you ask them to post in the forums here or contact us via the contact us option? https://secure.worldcommunitygrid.org/viewContactUs.do We are committed to being able to handle as large of user population as we can attract. Our current configuration is handling the current load very well (cpu use is only at 1-2% over long periods and IO throughput is holding up well also). We have additional servers that we are using for things such as beta testing that can be brought configuration to offload some of the work if server cpu usage becomes an issue. How many users are you planning on (or hoping) to attract to WCG? Your statements are giving me hopes of thousand's of new computers! |
||
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges:
|
After reading a bit on BOINC's website, the reasons for these work outages is becoming clearer. In an effort to make their client more attractive to DC project managers, berkeley has made the platform requirements very thin. So that anyone with "a single linux server" can host a new project. The problem lies in the fact that they do not require (or even encourage) higher levels of redundant hardware and infrastructure, so when the DB is flushed or a single PSU, HD, array etc, etc, etc goes down, so does the whole enchilada. Naturally, this makes starting a project alot easier and less expensive to implement and manage , however it makes the experience for the volunteer much more frustrating, which they try to counteract through a large variety of alternate, sometimes affiliated, projects. I disagree with the notion that there is anything inherently in BOINC that makes it prone to single points of failure or unscalable. For example, our database is backed up nightly and the disk storage for it is on network attached storage. We are in the process of building a server that is on standby that will serve as the failover for our database server. This server will simply mount the network attached storage, bring the database up and we are live again. We have monitors in place to detect outages and our response time to failures is quick. As far as scalability goes, all of the components of BOINC are designed so that multiple instances of each component (for example any of the server side dameons that load work, perform validation, etc) can be deployed onto multiple servers simultaneously. I highly suspect that the amount redundency and scalablity of any given project has more to do with the budget and time they have available to build a robust and redundant infrastructure. The comments about how you can run a BOINC project on a single server are there because that is in fact true and quite important to many research projects that have small budgets. However, it is also true that you can build a scalable and redundant infrastructure with BOINC. And given our friend Murphy - I'm sure we will have a failure in the next two minutes.  [Edit 3 times, last edit by knreed at Nov 4, 2005 5:00:45 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I would say take a look at einstein@home. Their project runs veery stable as well. They almost never had any issues and they give out as much work as other projects. It's just like kevin says, you can make boinc very stable with the right amount of fundings and time.
----------------------------------------http://einstein.phys.uwm.edu/ Edit: Look at the uptime of their server on the server status page The Einstein@Home main server has been continuously up for 126 days 16 hours 23 minutes. [Edit 1 times, last edit by Former Member at Nov 4, 2005 5:49:24 PM] |
||
|
|
|