Wed Jan 17 16:30:26 2007|Einstein@Home|Sending scheduler request to http://einstein.phys.uwm.edu/EinsteinAtHome_cgi/cgi
Wed Jan 17 16:30:26 2007|Einstein@Home|Reason: Requested by user
Wed Jan 17 16:30:26 2007|Einstein@Home|(not requesting new work or reporting completed tasks)
Wed Jan 17 16:30:32 2007|Einstein@Home|Scheduler request succeeded
Wed Jan 17 16:30:32 2007|Einstein@Home|Message from server: Project is temporarily shut down for maintenance
Wed Jan 17 16:30:32 2007|Einstein@Home|Project is down
This has been going on for some time. I find it difficult to believe that the server has been incapable of doling out work or accepting results for this period of time. I am not sure when the last time was that I did either, but I estimate two weeks.
As I alluded to in my previous answer, this is an entirely different issue to the current server problems. Even if the server were 100% operational at the moment, you would still get the "not requesting new work ..." message because it is coming directly from your BOINC client and not from the server. Change your mindset about the server being incapable of doling out work because this has nothing to do with your problem.
As I said last time, without significantly more information from you, anyone answering your question, including myself, is only guessing. Key bits of information needed are:-
*The version of BOINC you are using
*The projects you are supporting and their respective resource shares
*The value of your "connect to network every ..." preference setting
*Values of variables such as short_term_debt, long_term_debt, on_frac, connected_frac, active_frac, duration_correction_factor, and probably a few others that I've forgotten about at the moment :). All these can be examined in your state file "client_state.xml".
*Your crunching history over the last month, eg have any projects had to go into EDF mode as a result of deadline pressure, etc.
It is complex to write an answer explaining all possible factors and scenarios whilst it is actually reasonably easy to self diagnose, once you have a few basic principles of how the concept of debt works and how your BOINC client manages this. It does vary from version to version. The best place for information is undoubtedly the Wiki as a lot of time and effort has gone into explaining all these concepts there. Try searching for things like "work scheduler" or "work fetch policy", etc. Another thing to search on is the actual text of the message that is puzzling you, eg "Note: not requesting new work"
Finally, as a guess about the reasons for your problem, you probably have several projects and relatively small resource shares for some or all of these. You may also have low values for variables like on_frac and/or active_frac. You may also have an inappropriate value for your cache setting. EAH may have had far more than its fair share of your CPU at some stage (for whatever reason) and so the debt is now being repaid to other projects and BOINC is trying not to get into deadline trouble by preventing EAH from asking for more work until the debts are cleared.
EDIT: and I notice that Jord has covered a few other "shoot yourself in the foot" situations that could also be the reason behind this :). Thanks, Jord!!
I too have been unable to report any work or get any more. SETI is working fine, so I suspect is is something with Einstein. I am on a windows XP platform running Bionic 5.4.11. I am working on climateprediction.net, Einstein@Home, the almost defunct LHC@home and SETI@Home. All have 100%, so the have 25% of my system resources. I had a bad power supply, so haven't been crunching since Thanksgiving. I just got my rig back up and Einstein ran for several days fine, the nothing. I have been getting the same "Project is temporarily shut down for maintenance" since the 15th. I have the logs back that far if you want them. Here are just the last few entries.
1/17/2007 4:15:49 PM|climateprediction.net|Restarting task hadcm3ohe_0jk3_05673685_0 using hadcm3 version 515
1/17/2007 4:15:49 PM|SETI@home|Pausing task 23ap00aa.14262.28018.561076.3.205_1 (removed from memory)
1/17/2007 4:35:32 PM|Einstein@Home|Sending scheduler request to http://einstein.phys.uwm.edu/EinsteinAtHome_cgi/cgi
1/17/2007 4:35:32 PM|Einstein@Home|Reason: Requested by user
1/17/2007 4:35:32 PM|Einstein@Home|Requesting 43200 seconds of new work, and reporting 6 completed tasks
1/17/2007 4:35:37 PM|Einstein@Home|Scheduler request succeeded
1/17/2007 4:35:37 PM|Einstein@Home|Message from server: Project is temporarily shut down for maintenance
1/17/2007 4:35:37 PM|Einstein@Home|Project is down
1/17/2007 4:35:37 PM|Einstein@Home|Deferring scheduler requests for 2 hours, 5 minutes and 28 seconds
1/17/2007 5:15:49 PM|climateprediction.net|Pausing task hadcm3ohe_0jk3_05673685_0 (removed from memory)
1/17/2007 5:15:49 PM|SETI@home|Restarting task 23ap00aa.14262.28018.561076.3.205_1 using setiathome_enhanced version 515
1/17/2007 5:26:37 PM|SETI@home|Sending scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi
1/17/2007 5:26:37 PM|SETI@home|Reason: To fetch work
1/17/2007 5:26:37 PM|SETI@home|Requesting 5 seconds of new work
1/17/2007 5:26:42 PM|SETI@home|Scheduler request succeeded
1/17/2007 5:26:45 PM|SETI@home|Started download of file 30my00aa.5949.9728.1040906.3.152
1/17/2007 5:26:48 PM|SETI@home|Finished download of file 30my00aa.5949.9728.1040906.3.152
1/17/2007 5:26:48 PM|SETI@home|Throughput 121060 bytes/sec
1/17/2007 5:26:49 PM||Rescheduling CPU: files downloaded
For two–three days now, I have been sitting on three WUs ready to send, and my log keeps sending me this:
Thu 18 Jan 10:39:59 2007|Einstein@Home|Sending scheduler request to http://einstein.phys.uwm.edu/EinsteinAtHome_cgi/cgi
Thu 18 Jan 10:39:59 2007|Einstein@Home|Reason: Requested by user
Thu 18 Jan 10:39:59 2007|Einstein@Home|Reporting 3 tasks
Thu 18 Jan 10:40:04 2007|Einstein@Home|Scheduler request succeeded
Thu 18 Jan 10:40:04 2007|Einstein@Home|Message from server: Project is temporarily shut down for maintenance Thu 18 Jan 10:40:04 2007|Einstein@Home|Project is down
That means that maintenance is going on, and that the maintenance is unscheduled. It would be nice to receive just a one-liner in the news either acknowledging the problem, or assuring us that it will end.
George
------
Quote:
Dr George W Gerrity
4 Coral Place
Campbell, ACT 2612
AUSTRALIA
I too have been unable to report any work or get any more. SETI is working fine, so I suspect is is something with Einstein.
Try reading a few of the recent threads and you will see that everybody is in the same boat. Read Bernd's post (the oldest message) in the most recent sticky thread at the top of the "Cruncher's Corner" board. Reread this and other similar threads where the situation has been explained in answer to specific questions. I'm not about to write it all out yet again.
Quote:
I am on a windows XP platform running Bionic 5.4.11. I am working on climateprediction.net, Einstein@Home, the almost defunct LHC@home and SETI@Home. All have 100%, so the have 25% of my system resources. I had a bad power supply, so haven't been crunching since Thanksgiving. I just got my rig back up and Einstein ran for several days fine, the nothing. I have been getting the same "Project is temporarily shut down for maintenance" since the 15th. I have the logs back that far if you want them. Here are just the last few entries.
No need for any logs as everything is behaving as you would expect, given the server issues that are currently interfering with the distribution of new EAH work and the reporting of results. Because you have taken the sensible option of having multiple projects, Seti and CPDN will continue to keep your machine occupied. EAH and LHC will be accumulating positive debt which will be repaid to them at some stage in the future when these projects get more work. Under the circumstances, BOINC seems to be coping quite well with your situation. A perfect example of "Leave it alone and BOINC will sort it out for you".
You may notice an issue in the future. Because of the fact that you haven't been crunching for quite a while, BOINC will have noted this and will have reduced the value of the on_frac variable in your state file. So, future requests for work will be somewhat less than you might think, even if you leave your machine on 24/7, until the on_frac value recovers over time.
It would be nice to receive just a one-liner in the news either acknowledging the problem, or assuring us that it will end.
Every time you receive the log message you quoted, you are receiving a personal one-liner to inform you of the current status of the project. In other words you have received dozens of these over the last day or two. Why would a one-liner on the website be any more satisfying? Why do you need assurance that it will come to an end? Everything comes to an end at some point. There is nothing more certain than that.
I'm really not trying to be cute at all - just think about it a bit.
That fact that there are no further official pronouncements, apart from this one probably means that there is nothing further to add at this stage. In other words, the Devs are still grappling with the problem and don't have any useful information to give yet.
Please realise that Bernd is a programmer in Germany and probably not involved in day-to-day server issues. David Hammer looks after the servers but he may be away on holidays or just simply snowed under. If Bruce were able to, I'm sure he would have made some comment by now. He may be out of the country even. Take a look at his post count. He's not afraid to get his hands dirty.
Why do people seem to feel there is a conspiracy of silence? The much more plausible explanation is that Murphy has simply struck again. Things always go wrong at the most inconvenient time and in the most destructive manner.
Everybody just needs to curb the frustration and have a bit of patience.
Does this also affect current WU's that I have finished but haven't been able to upload to the server ?
I hope these will upload properly when the server is up and running again.
Andy
Many people seem to be wondering about much the same sort of things that you are. Some are even asking if they should abort the stuck results in their caches. I've just done a quick check of a number of my boxes. Based on that check, I estimate that I have more than 1500 results ready to report. I have no fears that eventually every last one of those will be reported and will be of use to the project.
From the behaviour of my own boxes, it's interesting to note that the ONLY things that are NOT happening at the moment are (i) the downloading of new work and (ii) the reporting of uploaded results. The contact with the scheduler is happening so that my boxes are not going into the 1 week backoff that some have reported. The maximum backoff I'm seeing is 2-3 hours and often a lot less. As results are finished, they are being uploaded - the status is always "Ready to report". The results are on the server waiting to be reported. Why the hell would anyone want to kill any results in this condition?
The following comments are not directed at any particular person. They are made simply to make sure people understand some basic issues.
Consider what would happen if any person decides to abandon any stuck results on their system. Since the quorum is two and the initial issue is two, you would be immdeiately forcing your quorum partner into an extended wait. All the time and effort you have put into the original crunching would be wasted and you would force the server to reissue the work to a third party.
So please don't walk away from EAH with no chance for the work to be reported. By all means go and crunch another project if you wish but leave EAH attached and unsuspended so that BOINC can clean up for you when the server is able to accept the reporting of your completed results. If you really want to abandon EAH just set "no new work" and finally detach when all your results have eventually been reported.
Gary,
please post THESE in the NEWS Public Region of EAH !
The one line on the main page does save a lot of panic. It is something that has always bugged me a bit with SETI - no news means bad news. I remember when SETI was out for almost a month and it turned out that a cut power line was the problem. How hard is it for some one to post:
*Your crunching history over the last month, eg have any projects had to go into EDF mode as a result of deadline pressure, etc.
Not aware of any similar problems.
It is complex to write an answer explaining all possible factors and scenarios whilst it is actually reasonably easy to self diagnose, once you have a few basic principles of how the concept of debt works and how your BOINC client manages this. It does vary from version to version. The best place for information is undoubtedly the Wiki as a lot of time and effort has gone into explaining all these concepts there. Try searching for things like "work scheduler" or "work fetch policy", etc. Another thing to search on is the actual text of the message that is puzzling you, eg "Note: not requesting new work"
Finally, as a guess about the reasons for your problem, you probably have several projects and relatively small resource shares for some or all of these. You may also have low values for variables like on_frac and/or active_frac. You may also have an inappropriate value for your cache setting. EAH may have had far more than its fair share of your CPU at some stage (for whatever reason) and so the debt is now being repaid to other projects and BOINC is trying not to get into deadline trouble by preventing EAH from asking for more work until the debts are cleared.
EDIT: and I notice that Jord has covered a few other "shoot yourself in the foot" situations that could also be the reason behind this :). Thanks, Jord!!
RE: Incidentally, I am
)
As I alluded to in my previous answer, this is an entirely different issue to the current server problems. Even if the server were 100% operational at the moment, you would still get the "not requesting new work ..." message because it is coming directly from your BOINC client and not from the server. Change your mindset about the server being incapable of doling out work because this has nothing to do with your problem.
As I said last time, without significantly more information from you, anyone answering your question, including myself, is only guessing. Key bits of information needed are:-
*The projects you are supporting and their respective resource shares
*The value of your "connect to network every ..." preference setting
*Values of variables such as short_term_debt, long_term_debt, on_frac, connected_frac, active_frac, duration_correction_factor, and probably a few others that I've forgotten about at the moment :). All these can be examined in your state file "client_state.xml".
*Your crunching history over the last month, eg have any projects had to go into EDF mode as a result of deadline pressure, etc.
It is complex to write an answer explaining all possible factors and scenarios whilst it is actually reasonably easy to self diagnose, once you have a few basic principles of how the concept of debt works and how your BOINC client manages this. It does vary from version to version. The best place for information is undoubtedly the Wiki as a lot of time and effort has gone into explaining all these concepts there. Try searching for things like "work scheduler" or "work fetch policy", etc. Another thing to search on is the actual text of the message that is puzzling you, eg "Note: not requesting new work"
Finally, as a guess about the reasons for your problem, you probably have several projects and relatively small resource shares for some or all of these. You may also have low values for variables like on_frac and/or active_frac. You may also have an inappropriate value for your cache setting. EAH may have had far more than its fair share of your CPU at some stage (for whatever reason) and so the debt is now being repaid to other projects and BOINC is trying not to get into deadline trouble by preventing EAH from asking for more work until the debts are cleared.
EDIT: and I notice that Jord has covered a few other "shoot yourself in the foot" situations that could also be the reason behind this :). Thanks, Jord!!
Cheers,
Gary.
I too have been unable to
)
I too have been unable to report any work or get any more. SETI is working fine, so I suspect is is something with Einstein. I am on a windows XP platform running Bionic 5.4.11. I am working on climateprediction.net, Einstein@Home, the almost defunct LHC@home and SETI@Home. All have 100%, so the have 25% of my system resources. I had a bad power supply, so haven't been crunching since Thanksgiving. I just got my rig back up and Einstein ran for several days fine, the nothing. I have been getting the same "Project is temporarily shut down for maintenance" since the 15th. I have the logs back that far if you want them. Here are just the last few entries.
1/17/2007 4:15:49 PM|climateprediction.net|Restarting task hadcm3ohe_0jk3_05673685_0 using hadcm3 version 515
1/17/2007 4:15:49 PM|SETI@home|Pausing task 23ap00aa.14262.28018.561076.3.205_1 (removed from memory)
1/17/2007 4:35:32 PM|Einstein@Home|Sending scheduler request to http://einstein.phys.uwm.edu/EinsteinAtHome_cgi/cgi
1/17/2007 4:35:32 PM|Einstein@Home|Reason: Requested by user
1/17/2007 4:35:32 PM|Einstein@Home|Requesting 43200 seconds of new work, and reporting 6 completed tasks
1/17/2007 4:35:37 PM|Einstein@Home|Scheduler request succeeded
1/17/2007 4:35:37 PM|Einstein@Home|Message from server: Project is temporarily shut down for maintenance
1/17/2007 4:35:37 PM|Einstein@Home|Project is down
1/17/2007 4:35:37 PM|Einstein@Home|Deferring scheduler requests for 2 hours, 5 minutes and 28 seconds
1/17/2007 5:15:49 PM|climateprediction.net|Pausing task hadcm3ohe_0jk3_05673685_0 (removed from memory)
1/17/2007 5:15:49 PM|SETI@home|Restarting task 23ap00aa.14262.28018.561076.3.205_1 using setiathome_enhanced version 515
1/17/2007 5:26:37 PM|SETI@home|Sending scheduler request to http://setiboinc.ssl.berkeley.edu/sah_cgi/cgi
1/17/2007 5:26:37 PM|SETI@home|Reason: To fetch work
1/17/2007 5:26:37 PM|SETI@home|Requesting 5 seconds of new work
1/17/2007 5:26:42 PM|SETI@home|Scheduler request succeeded
1/17/2007 5:26:45 PM|SETI@home|Started download of file 30my00aa.5949.9728.1040906.3.152
1/17/2007 5:26:48 PM|SETI@home|Finished download of file 30my00aa.5949.9728.1040906.3.152
1/17/2007 5:26:48 PM|SETI@home|Throughput 121060 bytes/sec
1/17/2007 5:26:49 PM||Rescheduling CPU: files downloaded
For two–three days now, I
)
For two–three days now, I have been sitting on three WUs ready to send, and my log keeps sending me this:
Thu 18 Jan 10:39:59 2007|Einstein@Home|Sending scheduler request to http://einstein.phys.uwm.edu/EinsteinAtHome_cgi/cgi
Thu 18 Jan 10:39:59 2007|Einstein@Home|Reason: Requested by user
Thu 18 Jan 10:39:59 2007|Einstein@Home|Reporting 3 tasks
Thu 18 Jan 10:40:04 2007|Einstein@Home|Scheduler request succeeded
Thu 18 Jan 10:40:04 2007|Einstein@Home|Message from server: Project is temporarily shut down for maintenance
Thu 18 Jan 10:40:04 2007|Einstein@Home|Project is down
That means that maintenance is going on, and that the maintenance is unscheduled. It would be nice to receive just a one-liner in the news either acknowledging the problem, or assuring us that it will end.
George
------
RE: I too have been unable
)
Try reading a few of the recent threads and you will see that everybody is in the same boat. Read Bernd's post (the oldest message) in the most recent sticky thread at the top of the "Cruncher's Corner" board. Reread this and other similar threads where the situation has been explained in answer to specific questions. I'm not about to write it all out yet again.
No need for any logs as everything is behaving as you would expect, given the server issues that are currently interfering with the distribution of new EAH work and the reporting of results. Because you have taken the sensible option of having multiple projects, Seti and CPDN will continue to keep your machine occupied. EAH and LHC will be accumulating positive debt which will be repaid to them at some stage in the future when these projects get more work. Under the circumstances, BOINC seems to be coping quite well with your situation. A perfect example of "Leave it alone and BOINC will sort it out for you".
You may notice an issue in the future. Because of the fact that you haven't been crunching for quite a while, BOINC will have noted this and will have reduced the value of the on_frac variable in your state file. So, future requests for work will be somewhat less than you might think, even if you leave your machine on 24/7, until the on_frac value recovers over time.
Cheers,
Gary.
RE: It would be nice to
)
Every time you receive the log message you quoted, you are receiving a personal one-liner to inform you of the current status of the project. In other words you have received dozens of these over the last day or two. Why would a one-liner on the website be any more satisfying? Why do you need assurance that it will come to an end? Everything comes to an end at some point. There is nothing more certain than that.
I'm really not trying to be cute at all - just think about it a bit.
That fact that there are no further official pronouncements, apart from this one probably means that there is nothing further to add at this stage. In other words, the Devs are still grappling with the problem and don't have any useful information to give yet.
Please realise that Bernd is a programmer in Germany and probably not involved in day-to-day server issues. David Hammer looks after the servers but he may be away on holidays or just simply snowed under. If Bruce were able to, I'm sure he would have made some comment by now. He may be out of the country even. Take a look at his post count. He's not afraid to get his hands dirty.
Why do people seem to feel there is a conspiracy of silence? The much more plausible explanation is that Murphy has simply struck again. Things always go wrong at the most inconvenient time and in the most destructive manner.
Everybody just needs to curb the frustration and have a bit of patience.
Cheers,
Gary.
RE: RE: RE: also see
)
Gary,
please post THESE in the NEWS Public Region of EAH !
Thanks Urban
http://www.boincstats.com/stats/banner.php?cpid=3837f9fafc28ff2e9df5b13ae2f8aaf7
The one line on the main page
)
The one line on the main page does save a lot of panic. It is something that has always bugged me a bit with SETI - no news means bad news. I remember when SETI was out for almost a month and it turned out that a cut power line was the problem. How hard is it for some one to post:
As of X date the project is down.
RE: *The version of BOINC
)
Also here is the info about
)
Also here is the info about my computer:
http://einsteinathome.org/host/721615
Also, I looked at the logs or
)
Also, I looked at the logs or my computer and this is what I discovered:
2007-01-19 01:53:43.7716 [PID=9636 ] [normal ] Handling request: host 721615, platform i686-apple-darwin, version 5.4.9, RSF 0.333333
2007-01-19 01:53:43.7717 [PID=9636 ] [normal ] OS version Darwin 8.8.1
2007-01-19 01:53:43.7774 [PID=9636 ] [debug ] Request [HOST#721615] Database [HOST#721615] Request [RPC#1250] Database [RPC#1249]
2007-01-19 01:53:43.7786 [PID=9636 ] [normal ] Processing request [HOST#721615] [RPC#1250] core client version 5.4.9
2007-01-19 01:53:43.8437 [PID=9636 ] [debug ] [HOST#721615] Resetting nresults_today
2007-01-19 01:53:43.8998 [PID=9636 ] [normal ] sending delay request 60.000000