Message boards :
Number crunching :
Almost out of Work...
Message board moderation
Author | Message |
---|---|
Jason Safoutin Send message Joined: 8 Sep 05 Posts: 1386 Credit: 200,389 RAC: 0 |
As you can see on the Server Status Page there are just over 70,000 WU's left to send. The updates are hours old in some cases and the splitters etc seem to be in the GREEN but we are almost out of work. Bad news is, it's Sunday...Will anyone be around to "fix" this (if needed)....or are we going to have to wait until Monday...Hmmm....*wonders* :-/ "By faith we understand that the universe was formed at God's command, so that what is seen was not made out of what was visible". Hebrews 11.3 |
W-K 666 Send message Joined: 18 May 99 Posts: 19012 Credit: 40,757,560 RAC: 67 |
|
Jakesnake5 Send message Joined: 14 May 99 Posts: 65 Credit: 18,370,396 RAC: 0 |
Look at the tape (splitter) positions. each tape holds 33000 blocks, each read is 48 blocks that are split into 256 WU's. It's logical that they wouldnt keep a couple million WU's ready to send you know. :) |
Swibby Bear Send message Joined: 1 Aug 01 Posts: 246 Credit: 7,945,093 RAC: 0 |
As you can see on the Server Status Page there are just over 70,000 WU's left to send. The updates are hours old in some cases and the splitters etc seem to be in the GREEN but we are almost out of work. Bad news is, it's Sunday...Will anyone be around to "fix" this (if needed)....or are we going to have to wait until Monday...Hmmm....*wonders* :-/ Read, Jason, Read! Matt said he might do an outage to again compress the db on Sunday or Monday. 70,000 is GREAT, considering that there were none available just a day or two ago. The splitters are doing fine, and there is nothing about the splitters that needs to be fixed. The deleters could sure use some help, though. The enhanced version will sure help all the processes when it is released. Whit |
Jack Gulley Send message Joined: 4 Mar 03 Posts: 423 Credit: 526,566 RAC: 0 |
I hope he runs into the lab today and does the compression and backup. The up/down/up/down of the Network link from the main site to the SSL lab is getting worse (been failing as of about midnight Friday night) and could go out at any time. Berkeley campus Network support will not even look at their "ticket" on this problem until 8:00 am Monday their time. And it could be at least four hours before they even start to look into the problem. If it is a Cogent link problem, then there could be a delay in getting them involved. So it would be best if he ran in today and started the process and just shut everything down now. Then hopefully when he is done after an over night run, the Network problem will be fixed and we can use the outage recovery as a test to see if this failing network hardware has been causing problem all along. |
Astro Send message Joined: 16 Apr 02 Posts: 8026 Credit: 600,015 RAC: 0 |
My best guess is that the DB issues are causing the slowness in getting work, the Splitters are probably more than capable of making work, but the data server is where the bottleneck is. The campus network is seperate from the Cogent link. This time it appears Cogent is working fine, but the SSL/campus network line is troubled. So, The Existing DB issues seem to still be with us, and the new Berkeley network line is having troubles. two seperate issues |
Swibby Bear Send message Joined: 1 Aug 01 Posts: 246 Credit: 7,945,093 RAC: 0 |
Actually, my thinking would be to wait until either the Berkeley router fails completely or the campus staff takes it down for repair/replacement. Then Matt can do the 3-hour or so outage to compress and reorganize the db while the line is down. That way, both problems will be worked on concurrently, rather than two separate outages for two different reasons. Matt, don'tcha love all the armchair advice? Whit |
Jakesnake5 Send message Joined: 14 May 99 Posts: 65 Credit: 18,370,396 RAC: 0 |
Having been the Admin for an ISP, I know how coordinating that would be. The IT guys wouldnt tell the lab guys WHEN their going to do work like that, so synchronizing would be a serious issue. Matt would have to be watching the network monitor just to GUESS it was down hard, and call IT to find out if their working on it. When I dealt with the providers here, it was worse than pulling hens teeth to get a straight answer from them. |
Mr.Pernod Send message Joined: 8 Feb 04 Posts: 350 Credit: 1,015,988 RAC: 0 |
let me do some simple math here. I average at around 17.6 granted credits per result (average over 270 results). SETI grants around 16~17M credits daily (see BOINCStats) This means they validate around 1M returned results per day. To replenish those, they need to create 1M results/day, which is close to a result creation rate of 11.6 results/second. The current creation rate, according to the serverstatus page, is around 12.8 results/second (1.1M results/day). If they can keep this up, they can even run up a few spare results to send out. |
John McLeod VII Send message Joined: 15 Jul 99 Posts: 24806 Credit: 790,712 RAC: 0 |
|
Paul D. Buck Send message Joined: 19 Jul 00 Posts: 3898 Credit: 1,158,042 RAC: 0 |
But, the data rate of the telescope is only about 1 Million results per day. Well, Matt still has a few spare tapes built up from the past ... |
Mr.Pernod Send message Joined: 8 Feb 04 Posts: 350 Credit: 1,015,988 RAC: 0 |
But, the data rate of the telescope is only about 1 Million results per day. The most recent tape I have seen being split so far dates back to march 2005, so I think there's plenty of work waiting in the closet to keep us fed for a long time, even if there a lot of noisy results and maybe even a couple of blank tapes. Besides, once Enhanced hits, they better get some more closetspace to store the new tapes ;) |
Jason Safoutin Send message Joined: 8 Sep 05 Posts: 1386 Credit: 200,389 RAC: 0 |
But, the data rate of the telescope is only about 1 Million results per day. I just crunched a WU yesterday from I think march of 03. "By faith we understand that the universe was formed at God's command, so that what is seen was not made out of what was visible". Hebrews 11.3 |
Matt Lebofsky Send message Joined: 1 Mar 99 Posts: 1444 Credit: 957,058 RAC: 0 |
The reason the most recent tape was split was around March 05 is not because we have a huge backlog of data - it's because we couldn't afford any more DLT tapes and stopped recording data around that time. Plus all our old DLT drives are broken - we have only one left up here at the lab that reasonably works - that's what I use to read these old data tapes. I think I have about 50 tapes left in my pile to split. At the current "burn rate" this should last about a month. However! We are releasing SETI enhanced, which should slow everything down, and this current set of data should last many more months - hopefully long enough to get the new data recorder down to Arecibo and get that all working. And! Due to one reason or another we do have many tapes that were split and sent out but we have only about 10% scientific return - so we may throw those back into the pipeline if we otherwise run out of workunits to send out. - Matt -- BOINC/SETI@home network/web/science/development person -- "Any idiot can have a good idea. What is hard is to do it." - Jeanne-Claude |
Mr.Pernod Send message Joined: 8 Feb 04 Posts: 350 Credit: 1,015,988 RAC: 0 |
Taking into account the higher sensitivity of the upcoming Enhanced Application, why not recycle some old tapes with a "large" number of likely candidates? (assuming you kept the old tapes) |
kevint Send message Joined: 17 May 99 Posts: 414 Credit: 11,680,240 RAC: 0 |
. Plus all our old DLT drives are broken - we have only one left up here at the lab that reasonably works - that's what I use to read these old data tapes. Matt - I have a Compaq DLT 40/80 series 3306 unit - it is not doing anything but collecting dust here - How can I get it to you ? |
MJKelleher Send message Joined: 1 Jul 99 Posts: 2048 Credit: 1,575,401 RAC: 0 |
From Donate Hardware to SETI@Home:If you have any interest in making such a donation, please send questions/comments to: donors@ssl.berkeley.edu.. Plus all our old DLT drives are broken - we have only one left up here at the lab that reasonably works - that's what I use to read these old data tapes. MJ |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.