Message boards :
Number crunching :
Deadline Differences between Website and Client
Message board moderation
Author | Message |
---|---|
Send message Joined: 2 Sep 04 Posts: 453 Credit: 193,576,736 RAC: 5,548 |
|
Send message Joined: 28 Sep 04 Posts: 677 Credit: 43,759,057 RAC: 14,785 |
I remember this being discussed in the forums several years ago. No remedy was made about it. I couldn't find the message thread in question. |
Send message Joined: 15 Jun 08 Posts: 2425 Credit: 227,425,457 RAC: 130,174 |
CMS deadlines also: Local: 2017-04-09:04:41:08 CEST Server: 2017-04-10:02:41:08 UTC The clocks will shortly change to DST but this does not happen at the same date around the world. It does not explain a one day difference but it may point to a general bug in the time calculation. |
Send message Joined: 2 May 07 Posts: 2120 Credit: 159,926,969 RAC: 70,085 |
The same procedure as last year! The CEST is correct for those tasks with end-time after the date when CEST begin. Edit: you see the same with one hour less at the end of Octobre. |
Send message Joined: 2 Sep 04 Posts: 453 Credit: 193,576,736 RAC: 5,548 |
|
Send message Joined: 15 Jun 08 Posts: 2425 Credit: 227,425,457 RAC: 130,174 |
It may skip a whole day due to a bug. Let´s crosscheck this with other projects. At the moment I have no WUs with deadlines after clock change (except CMS) and all of them show normal time calculation. If the calculations show errors in a few days and return back to normal after the clock change then it´s related to DST. If the error occurs only with LHC then it is a server problem, otherwise it´s unclear. |
Send message Joined: 15 Jun 08 Posts: 2425 Credit: 227,425,457 RAC: 130,174 |
Meanwhile I got a couple of WUs from other projects with due dates after the clock change. All of them show correct time calculations. The problem seems to be CERN specific. |
Send message Joined: 15 Jun 08 Posts: 2425 Credit: 227,425,457 RAC: 130,174 |
Meanwhile I got a couple of WUs from other projects ...... including ATLAS (unconsolidated server). The anomaly persists for CMS (only). |
Send message Joined: 2 Sep 04 Posts: 453 Credit: 193,576,736 RAC: 5,548 |
|
Send message Joined: 14 Jan 10 Posts: 1286 Credit: 8,515,710 RAC: 2,852 |
I've seen that sometimes in earlier days with other projects too. The purpose is/was that the client will start running in high priority on time and is still able to return on time (24hrs spare time left) and a resend task from the project is not needed. |
Send message Joined: 15 Jun 08 Posts: 2425 Credit: 227,425,457 RAC: 130,174 |
@ Yeti ATLAS tasks from unconsolidated (1) or consolidated (2) server? Mine are from (1) as I do not (yet) run ATLAS from (2). @ CP I run a small buffer so my CMS tasks start immediately after they are downloaded. They are far away from the 24h border. |
Send message Joined: 2 Sep 04 Posts: 453 Credit: 193,576,736 RAC: 5,548 |
|
©2024 CERN