Message boards :
Number crunching :
Missing Results?
Message board moderation
Author | Message |
---|---|
Rasputin42 Send message Joined: 25 Jul 08 Posts: 412 Credit: 5,834,661 RAC: 0 |
Hi, I cannot seem to find any results validated or pending for the day of the outage (11 aug UTC) and maybe half a day before. Has anyone noticed the same? |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
All mine look normal, except of course there's a big gap between ~15:00 UTC (early start to the outage) and the early hours of 12 August (late finish). There are problems with the workunit pages for both MB and AP, as we've discussed elsewhere, but not result pages. Unless you have information from your logs that suggest otherwise? |
Rasputin42 Send message Joined: 25 Jul 08 Posts: 412 Credit: 5,834,661 RAC: 0 |
I have no results, that seem to have been received by the server from about 5pm aug 10 to the end of aug 11 (UTC time). I am not 100% certain, but there had to be results turned in by my computer during that time. I noticed a sharp drop in my RAC beforethe outage. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
I'd suggest that you check things at your end. Looking in particular at your Computer 7593028, you have loads of work available to work on, but you've returned very little of it in the last two or three weeks - 25 of the tasks passed their deadlines without being processed. That looks from here more like the computer wasn't running BOINC, or wasn't switched on, even though you may have thought it was. Although the project has had one or two problems, there's been nothing that could explain a gap as long as that - and your computer should have been able to continue processing the work it had available, even if the project was down. |
Rasputin42 Send message Joined: 25 Jul 08 Posts: 412 Credit: 5,834,661 RAC: 0 |
Thanks. The timed out ones were lost units from a few weeks ago. Maybe it is nothing, but my computer was working all the time. I just never had a whole day, where nothing was reported (while running) Working the time back, from the first results reported on the 12 aug utc, there should have been a few results arriving at the server before the outage. Thanks for looking into that, maybe i am a just seeing things(or not seeing, in this case) |
rob smith Send message Joined: 7 Mar 03 Posts: 22202 Credit: 416,307,556 RAC: 380 |
Don't forget that any result that has been validated will vanish from your list after 24 hours. So it is possible that you have returned results where you wingman was "waiting" for you, in which case validation would be "instant, and then 24 hours later they don't show... Bob Smith Member of Seti PIPPS (Pluto is a Planet Protest Society) Somewhere in the (un)known Universe? |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
Don't forget that any result that has been validated will vanish from your list after 24 hours. So it is possible that you have returned results where you wingman was "waiting" for you, in which case validation would be "instant, and then 24 hours later they don't show... That's certainly the case for Rasputin, because he's currently working on tasks issued om 31 July 2015, before we reached task 2^32. It's even possible that in their efforts to clean up the database last night, the crew ran an extra DB purge which removed validated tasks earlier than the normal 24 hour grace period. Fir the rest of us, working on tasks issued after Task 2^32 (2 Aug 2015, 5:44:59 UTC), the normal 'disappears 24 hours after validation' rule doesn't currently apply. |
Zombu2 Send message Joined: 24 Feb 01 Posts: 1615 Credit: 49,315,423 RAC: 0 |
Hmm it s weird i am cranking out wu's but they do not show up as pending or validated they disappear into thin air I came down with a bad case of i don't give a crap |
Rasputin42 Send message Joined: 25 Jul 08 Posts: 412 Credit: 5,834,661 RAC: 0 |
the normal 'disappears 24 hours after validation' rule doesn't currently apply. What does, instead? |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
Don't forget that any result that has been validated will vanish from your list after 24 hours. So it is possible that you have returned results where you wingman was "waiting" for you, in which case validation would be "instant, and then 24 hours later they don't show... I have several, like this wu, that were issues & completed just before maintenance started. So I'm not sure if they ran an early purge. Since the wu will be purged soon. Here are the timestamps of the two results. Sent: 11 Aug 2015, 10:41:43 UTC, Reported: 11 Aug 2015, 14:41:22 UTC Sent: 11 Aug 2015, 10:41:46 UTC, Reported: 11 Aug 2015, 12:18:40 UTC Perhaps that one didn't make it through the validator stage before maintenance started, but I have dozens that were reported & validated prior to maintenance. The oldest wu I can see with both results now goes back to 11 Aug 2015, 13:13:01 UTC. As the rest have started the delete/purge process. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
the normal 'disappears 24 hours after validation' rule doesn't currently apply. Certainly up until yesterday, tasks with TaskID >= 2^32 weren't being purged at all - they were sill visible in the database 10 days after validation. That's related to the problem of 12 million (and rising) undeleted result files shown on the server status page. Certainly, if I were a server admin, my first priority would be to fix the purge/deleter process so it dealt with 64-bit numbers, so that the problem doesn't get any worse. Maybe they reduced the purge delay, so that they could see whether it was working for new validations as soon as they get into the lab today, rather than waiting 24 hours until practically the end of the working day today. That might account for Zombu's observation. Then they would have capped the problem, and they would work on the backlog at leisure, and make sure they do it right. |
Rasputin42 Send message Joined: 25 Jul 08 Posts: 412 Credit: 5,834,661 RAC: 0 |
Thanks, Guys! I just wanted to know, if there was something odd going on. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
Watching Workunit 1865970821 4310560648 6910484 11 Aug 2015, 12:14:01 UTC 11 Aug 2015, 14:34:39 UTC Completed and validated 3,246.83 3,223.07 489.70 AstroPulse v7 Anonymous platform (NVIDIA GPU) 4310560649 6946318 11 Aug 2015, 12:14:01 UTC 11 Aug 2015, 13:36:56 UTC Completed and validated 2,575.00 436.38 489.70 AstroPulse v7 v7.07 (opencl_nvidia_mac) According to theory, the whole caboodle (WU and results) should be purged in about an hour... |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
No such luck. Task 4310560648 is still in the database, but the associated WU 1865970821 was purged after 24 hours as usual. So the pile of un-purged results is going to get worse. |
arkayn Send message Joined: 14 May 99 Posts: 4438 Credit: 55,006,323 RAC: 0 |
Watching And work unit gone. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
And work unit gone. Leaving both results orphaned. |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
No such luck. Task 4310560648 is still in the database, but the associated WU 1865970821 was purged after 24 hours as usual. Indeed every task from 4294967296 to 4294967250 is still present in the system with its associated workunit having been purged. While tasks from 4294967295 to 4294967250 & earlier have been purged. The newest task I have is 4312462714. So that gives us at least 17,495,418 tasks that will be orphaned. So the current total of 12,714,491 Results waiting for deletion doesn't even account for all of the new work generated. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14650 Credit: 200,643,578 RAC: 874 |
No such luck. Task 4310560648 is still in the database, but the associated WU 1865970821 was purged after 24 hours as usual. And, of course, the BOINC database is shared between MB and AP, so recent AP WUs (like the mine-canary I picked) have IDs in the vulnerable range. At least the AP heap started smaller, and grows more slowly - but it's nonetheless a heap, and it looks like we'll be adding to the MB heap again soon. |
©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.