Forums :
Technical Support :
URGENT Problems Discussion Thread
Message board moderation
Previous · 1 . . . 12 · 13 · 14 · 15 · 16 · 17 · 18 · Next
Author | Message |
---|---|
![]() Volunteer moderator Project administrator Project developer ![]() Send message Joined: 1 Apr 07 Posts: 662 Credit: 13,742 RAC: 0 |
Over the last couple of hours, the number of WUs in progress has about doubled, so I think we\'re making progress. Scott Kruger Project Administrator, Cosmology@Home |
STE\/E Volunteer tester Send message Joined: 12 Jun 07 Posts: 375 Credit: 16,539,257 RAC: 0 |
Over the last couple of hours, the number of Wu\'s in progress has about doubled, so I think we\'re making progress. Yes, I\'ve managed to Download quite a few Wu\'s, enough to keep my PC\'s busy for a little while anyway. Problem is now the Server is acting slow again making it hard to navigate around the Forums & Accounts, also Uploading works seems to be okay but Reporting them seems to take quite awhile ... !!! |
![]() ![]() Send message Joined: 28 Aug 07 Posts: 169 Credit: 1,587,686 RAC: 1,370 ![]() |
I have almost run out on my Windows machine (AMD Opteron 285), only got another half dozen or so, which will soon be gone. One Linux machine got a queue full and is now happy, one got about 6 and now empty again, and one Linux machine has had no work for two days (all AMD Opteron). I am now getting the \"work for other platforms\" message which I did not get before. Just checked and found the machine that has had no work for 2 days, has in fact downloaded some work but it was like Poorboy is said, they were all MD5 checksum errors so no work was actually processed. |
Fred Send message Joined: 17 Jan 08 Posts: 40 Credit: 228,230 RAC: 0 |
Likewise on my Intel box (Q6600/XP Home) I am getting only: Sorry for the bad form of replying to my own post but this is really an extension of the previous comment. I continued to get the same response to every request for new work (i.e. just a \"no work from project\" without anything else) until my supply dried up totally - none left. At that point, I detached and reattached and, hey presto, new work downloaded :) This may be a way to \"kick the server\" if you run dry? F. ![]() |
![]() ![]() Send message Joined: 19 Dec 07 Posts: 24 Credit: 889,050 RAC: 0 |
Hmm, I think you just got lucky. Just tried that on each host as they ran dry to no avail. :( They all say: Mon 07 Apr 2008 04:01:35 AM HST|Cosmology@Home|Reason: no work from project Too bad cause it was fun while it lasted. Everything was running great including the website. Now the site is crawling, no errors yet though. Uploads seem to have slowed to a crawl as well. Currently the server status page shows green across the board: Results ready to send 77,771 Results in progress 88,556 waiting for validation 23 ![]() |
Fred Send message Joined: 17 Jan 08 Posts: 40 Credit: 228,230 RAC: 0 |
Maybe the difference is that I was getting only \"Scheduler request succeeded: got 0 new tasks\". No explanation from the Server. You are getting \"Reason: no work from project\" so, at least the server is talking to your host. I am coming to the conclusion that the server status page is just a pretty picture with some random numbers thrown in to make it more interesting - a figment of a deranged imagination - and bears no resemblance to the status of the servers at all. [whinge]And, yes - navigating these pages is like wading through treacle again!![/whinge] F. ![]() |
![]() Send message Joined: 30 Aug 07 Posts: 46 Credit: 6,502,980 RAC: 0 |
Not being able to report again. Uploads are slow, if not stopped. site is slow. Seems like Monday morning. Hope we don\'t have to wait till Friday when Scott gets done with his classes before we can get this corrected. Maybe a grand idea would be to just not issue any more work, let the caches on all boxes clear, validate all work. Fix the problem. Then start to release work again. Basically start out at zero work in progress.... |
STE\/E Volunteer tester Send message Joined: 12 Jun 07 Posts: 375 Credit: 16,539,257 RAC: 0 |
[whinge]And, yes - navigating these pages is like wading through treacle again!![/whinge] F You got that right, you click on Post Reply then settle in for a Nap ... hahaha ... Hopefully by the time your Naps done the Page has opened ... ;) Anyway I\'ve managed to Download about 3500 Wu\'s across my Farm so work can be downloaded if your Persistent enough. I\'m running Winblows XP Pro 64 Bit on Intel Quads > The Q6600 & Q6700 Vintage, I don\'t know about other CPU\'s or OS\'s if they can get work or not. Getting work is one thing, Uploading finished Wu\'s doesn\'t seem bad most of the time unless Scott has the Server down. But Returning them is a whole new ballgame, at time it\'s not a problem, other times you have to wait an hour or longer to get them Returned ... :) |
![]() Volunteer moderator Project administrator Project developer ![]() Send message Joined: 1 Apr 07 Posts: 662 Credit: 13,742 RAC: 0 |
I have to restart the daemons and the database every once in a while because at some point, queries just stop wanting to finish. After the restart, everything works fine for a while, but eventually things get all slow again. This problem is really elusive. I\'ll have to locate a mysql expert to help me out. Scott Kruger Project Administrator, Cosmology@Home |
![]() Volunteer moderator Project administrator Project developer ![]() Send message Joined: 1 Apr 07 Posts: 662 Credit: 13,742 RAC: 0 |
I made some changes to the database which I think should do some good. Only time will tell though... If you guys notice that the site is faster over the next couple of hours, please say so. Scott Kruger Project Administrator, Cosmology@Home |
![]() Send message Joined: 30 Aug 07 Posts: 46 Credit: 6,502,980 RAC: 0 |
I made some changes to the database which I think should do some good. Only time will tell though... The web site seems faster- right now. But maybe that is because the schedulers are down and no work is being uploaded / downloaded or reported. I also noticed you are lowering credit again. Does this affect all newly release work, or work that is already cached. |
![]() Volunteer moderator Project administrator Project developer ![]() Send message Joined: 1 Apr 07 Posts: 662 Credit: 13,742 RAC: 0 |
I also noticed you are lowering credit again. Does this affect all newly release work, or work that is already cached. Newly-released work only. Scott Kruger Project Administrator, Cosmology@Home |
![]() Volunteer moderator Volunteer tester ![]() Send message Joined: 7 Jun 07 Posts: 217 Credit: 710,406 RAC: 0 |
I made some changes to the database which I think should do some good. Only time will tell though... Some 6 hours after your action. The site seems faster at times, but it still has slow periods. |
Fred Send message Joined: 17 Jan 08 Posts: 40 Credit: 228,230 RAC: 0 |
I made some changes to the database which I think should do some good. Only time will tell though... I would put that comment in the reverse order from where I am sitting: Still slow navigating the web pages but with occasional flashes of inspiration. F. ![]() |
![]() ![]() Send message Joined: 28 Aug 07 Posts: 169 Credit: 1,587,686 RAC: 1,370 ![]() |
Got the message \"there was work for other platforms\" as of a few hours ago, but this does not appear to be stopping me getting plenty of work on all computers, both Linux and Windows. The site is quite a bit faster at the moment compared to 4 hours ago, I have actually been able to check my pendings and check my computers, albeit slowly. Overall the site is functional and thanks Scott, for keeping us in the loop. |
![]() Send message Joined: 1 Jul 07 Posts: 37 Credit: 208,284 RAC: 0 |
I would agree with Fred. Also am now getting these messages so all is not well. 08/04/2008 13:21:12|Cosmology@Home|Message from server: Server can\'t open database 08/04/2008 13:27:14|Cosmology@Home|Message from server: No work sent Dave. |
STE\/E Volunteer tester Send message Joined: 12 Jun 07 Posts: 375 Credit: 16,539,257 RAC: 0 |
It\'s been really Slooooooowwwwwwwww for me all morning, it took me several hours to Return a few Hundred Finished Wu\'s that built up while I slept. I had a Root Canal done a few weeks ago that was less Painful than trying to access or once accessed to maneuver around the Site ... ;) ... I haven\'t been able to see my Pending Wu\'s for a few days either. |
Phoneman1 Send message Joined: 5 Nov 07 Posts: 113 Credit: 3,100,327 RAC: 0 |
I agree with the last three posts. One of my Intel quads has 4 work units running and a queue of 5 more. The second machine is out of work and has been for a few hours. I\'ve put it on to another project for now. When I checked the site a couple of hours ago I couldn\'t even login. I got the \"1040Too many connections\" from mySQL. Although it is possible for database administrators to increase the number of connections this is sticky plaster action really and I wouldn\'t recommmend it just now. I think the reason for the message is that database accesses are all taking longer than they should. In turn this means there are more concurrent accesses taking place. As this database is getting a lot more use with the shorter work units I suspect that its records and indexes are occupying \"overflow pages\" more than usual. I\'ve seen just this sort of effect we are experiencing now on a DB2 database where this happened. I don\'t know what stats Scott is able to get from mySQL re the use of \"overflow pages\" by the system or what mySQL offers to correct the problem but it is one line of enquiry that could be worth looking into. Chances are mySQL calls \"overflow pages\" something else but I strongly suspect the problem is to do with the rate at which database records and their associated indexes are now being added, updated and deleted which leads to them being put in the \"wrong place\" and subsequent accesses are required to find them. Phoneman1 |
![]() Volunteer moderator Project administrator Project developer ![]() Send message Joined: 1 Apr 07 Posts: 662 Credit: 13,742 RAC: 0 |
I checked the memory usage this morning and it looks like when mysql eventually uses all of the memory buffer pool, it\'s taking up about half the server memory. This wouldn\'t be an issue except for that, somehow, the other half of the memory is also being used for some reason. This is most likely causing some serious paging issues and the slowdown. So, this is the next thing I\'ll check out. Scott Kruger Project Administrator, Cosmology@Home |
Brian Silvers Send message Joined: 11 Dec 07 Posts: 420 Credit: 270,580 RAC: 0 |
Validator is kaput. Server status page is kaput... Reporting did work though... ![]() |