Message boards :
Number crunching :
Message from server: Data collect is not available for your type of computer
Message board moderation
Author | Message |
---|---|
Send message Joined: 10 May 10 Posts: 15 Credit: 55,797 RAC: 0 |
Continuing the News : New application released thread here: Thyme Lawn wrote:
This might be a reason, but then the complaint should be worded pretty much differently) like "This computer has reached a limit on consecutive failing tasks"), shouldn't it? My host 3973 did not have any failed tasks, at least recently - the last 16 results (3 days) kept in list (each "Data collect version 3 v3.16 (nci)") were all validated and credited. But still I've noticed the same notice and later a bunch of these errors in Messages: 04.07.2011 7:47:21 | WUProp@Home | Sending scheduler request: Requested by project. 04.07.2011 7:47:21 | WUProp@Home | Not reporting or requesting tasks 04.07.2011 7:47:25 | WUProp@Home | Scheduler request completed --- 04.07.2011 8:47:25 | WUProp@Home | Sending scheduler request: Requested by project. 04.07.2011 8:47:25 | WUProp@Home | Not reporting or requesting tasks 04.07.2011 8:47:27 | WUProp@Home | Scheduler request completed 04.07.2011 8:47:27 | WUProp@Home | Server can't open database 04.07.2011 8:47:29 | WUProp@Home | Computation for task wu_v3_1308579301_383167_0 finished --- 04.07.2011 9:47:30 | WUProp@Home | Sending scheduler request: Requested by project. 04.07.2011 9:47:30 | WUProp@Home | Reporting 1 completed tasks, requesting new tasks for CPU 04.07.2011 9:47:33 | WUProp@Home | Scheduler request completed: got 0 new tasks 04.07.2011 9:47:33 | WUProp@Home | Server can't open database --- 04.07.2011 10:47:35 | WUProp@Home | Sending scheduler request: Requested by project. 04.07.2011 10:47:35 | WUProp@Home | Reporting 1 completed tasks, requesting new tasks for CPU 04.07.2011 10:47:43 | WUProp@Home | Scheduler request completed: got 0 new tasks 04.07.2011 10:47:43 | WUProp@Home | No work sent 04.07.2011 10:47:43 | WUProp@Home | Message from server: Data collect nie je dostupný pre váš typ poÄÃtaÄa. - 04.07.2011 10:47:53 | WUProp@Home | Sending scheduler request: To fetch work. 04.07.2011 10:47:53 | WUProp@Home | Requesting new tasks for CPU 04.07.2011 10:48:13 | WUProp@Home | Scheduler request completed: got 0 new tasks 04.07.2011 10:48:13 | WUProp@Home | No work sent 04.07.2011 10:48:13 | WUProp@Home | Message from server: Data collect nie je dostupný pre váš typ poÄÃtaÄa. --- 04.07.2011 10:57:23 | WUProp@Home | Sending scheduler request: To fetch work. 04.07.2011 10:57:23 | WUProp@Home | Requesting new tasks for CPU 04.07.2011 10:57:27 | WUProp@Home | Scheduler request completed: got 1 new tasks 04.07.2011 10:57:29 | WUProp@Home | Starting task wu_v3_1308579301_389110_0 using data_collect_v3 version 316 --- 04.07.2011 13:57:37 | WUProp@Home | Computation for task wu_v3_1308579301_389110_0 finished 04.07.2011 13:57:39 | WUProp@Home | Started upload of wu_v3_1308579301_389110_0_0 04.07.2011 13:57:44 | WUProp@Home | Sending scheduler request: Requested by project. 04.07.2011 13:57:44 | WUProp@Home | Requesting new tasks for CPU 04.07.2011 13:57:49 | WUProp@Home | Finished upload of wu_v3_1308579301_389110_0_0 04.07.2011 13:57:49 | WUProp@Home | Scheduler request completed: got 0 new tasks 04.07.2011 13:57:49 | WUProp@Home | No work sent 04.07.2011 13:57:49 | WUProp@Home | This computer has reached a limit on tasks in progress - 04.07.2011 13:58:00 | WUProp@Home | Sending scheduler request: To fetch work. 04.07.2011 13:58:00 | WUProp@Home | Reporting 1 completed tasks, requesting new tasks for CPU 04.07.2011 13:58:18 | WUProp@Home | Scheduler request completed: got 1 new tasks 04.07.2011 13:58:21 | WUProp@Home | Starting task wu_v3_1308579301_392816_0 using data_collect_v3 version 316 According to the log, I rather assume some tinkering with the server - initially all was fine, then "Server can't open database", then "No work sent, Data collect is not available for your type of computer", again once more, then suddenly "got 1 new tasks". Was this also the Nosferatu's case? One unfortunate "resources contention" is still happening in 6.12.33 (13:57:37 - 13:57:49): the client would like to start crunching as soon as possible (right after finishing the previous task), but the server is not willing to deliver a new one prior to uploading the finished one. Thus the client requests a new task just during an upload - I do not remember whether doing a scheduler request during an upload has an exception, if no tasks are available... Peter PS: I do HATE Akismet!!! (Will this trigger it too?) I was at least able to inject all links afterwards. |
Send message Joined: 11 Apr 10 Posts: 54 Credit: 382,341 RAC: 0 |
Pepo wrote: This might be a reason, but then the complaint should be worded pretty much differently) like "This computer has reached a limit on consecutive failing tasks"), shouldn't it? I beg to differ Peter. There are circumstances where a "consecutive failing tasks" message would be totally misleading (e.g. if a project doesn't have much work available it might set the daily quota lower than the number of tasks a CPU core can complete in a day to give all users a decent chance of getting work). Pepo wrote: According to the log, I rather assume some tinkering with the server - initially all was fine, then "Server can't open database", then "No work sent, Data collect is not available for your type of computer", again once more, then suddenly "got 1 new tasks". Was this also the Nosferatu's case? No. Nosferatu's messages included "This computer has finished a daily quota of 1 tasks" which is missing from yours. The "not available" message were probably a legacy of the database having been down. Possibly the feeder queue was still empty. Pepo wrote: I do not remember whether doing a scheduler request during an upload has an exception, if no tasks are available... Scheduler requests include a list of in progress tasks in the <other_results> block which only excludes tasks which are being reported. Excluding an uploading task from that section would result in the server resending it (if it's configured to resend lost tasks). Perhaps there's an argument for an <uploading/> tag in <other_result> sections as an indication to (in particular) NCI projects that a task shouldn't be counted as in progress when work is being requested. |
©2024 Sébastien