Message boards :
Number crunching :
Purging Results
Message board moderation
Author | Message |
---|---|
Send message Joined: 24 Dec 07 Posts: 1947 Credit: 240,884,648 RAC: 0 |
From the front page....I've had to start purging results immediately (as opposed to leaving them on the server for a day) because our server has been getting overloaded with the new faster applications out there. Bugger. Be nice if they could hang around for a few hours. I use my task results to help see how my computers are running. Any chance of finding a happy medium? |
Send message Joined: 4 Oct 08 Posts: 1734 Credit: 64,228,409 RAC: 0 |
Heck! They are being purged fast. No records of performance or success/client error. I can only assume things are running fine and results are correct/acceptable, as it has these last few weeks. |
Send message Joined: 12 Nov 07 Posts: 2425 Credit: 524,164 RAC: 0 |
On the 14th after that was posted the results seemed to be around for a few hours, and that got shorter to the current under 15 min. (from what I've noticed) Doesn't expecting the unexpected make the unexpected the expected? If it makes sense, DON'T do it. |
Send message Joined: 12 Nov 07 Posts: 2425 Credit: 524,164 RAC: 0 |
When the new app is released, can the wu results be left for a longer period of time (6-12 hours) so that they can be checked to see that everything works? Doesn't expecting the unexpected make the unexpected the expected? If it makes sense, DON'T do it. |
Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
Yes, we'll be leaving results around longer with the new version of the app. I'm going to try to recode db_purge so we can purge results every X hours, instead of every X days (which is the way it is now). But generally, our assimilator writes out the results in our search files, so we know if you're work is good or not, so there's really not much reason to leave them in the database slowing things down (at least on our side of things). |
Send message Joined: 12 Nov 07 Posts: 2425 Credit: 524,164 RAC: 0 |
I still like to see that the results have gone through correctly and my computer didn't go funny. Doesn't expecting the unexpected make the unexpected the expected? If it makes sense, DON'T do it. |
Send message Joined: 7 Jun 08 Posts: 464 Credit: 56,639,936 RAC: 0 |
Also, some of us like to log the task output, and unfortunately there is no mechanism in BOINC to relay the TID/WUID or granted credit for individual tasks back to the CC. Therefore, the only way to get that is to periodically examine the computer summary pages. With the stock apps the quick purge hasn't been much of a problem except for the TID/RID, but with the third party apps it's more problematical. Alinator |
Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 |
We've been having some pretty big problems with the NFS system milkyway is on at RPI. Milkyway has been using up more than it's own hard drive space so it's been spilling over onto NFS which has really slowed things down. We're ordering a new group of RAID hard drives for milkyway which should help alleviate that problem, but until then we're gonna have to clear em as fast as we get them. |
Send message Joined: 12 Nov 07 Posts: 2425 Credit: 524,164 RAC: 0 |
How long are the test results going to be pending? Will they be validated? Doesn't expecting the unexpected make the unexpected the expected? If it makes sense, DON'T do it. |
Send message Joined: 12 Nov 07 Posts: 2425 Credit: 524,164 RAC: 0 |
It needs to be changed to atleast a few hours if not 24. Especially if you want anyone to give refrences to current credit. Doesn't expecting the unexpected make the unexpected the expected? If it makes sense, DON'T do it. |
Send message Joined: 10 Aug 08 Posts: 218 Credit: 41,846,854 RAC: 0 |
It needs to be changed to atleast a few hours if not 24. Especially if you want anyone to give refrences to current credit... Or errors. I'm getting some errors here with the latest client and its gone before I can examine it to post the results. |
©2024 Astroinformatics Group