 |
 |
Hi, This is because of the way how SFT2 is publishing the results. After several discussions we decided that we want a mechanism that allows to see results as quickly as possible even if some tests have not finished yet. That is why currently all tests are independed and run in parallel, although being part of a single test job. After each single test finishes its results are immediately published (using internal SFT transport mechanism). As for job submission itself, if there is ?? for job submission and the other tests are OK it simply means, that the test job started, the tests inside were executed successfully and they published the results but the system is still processing the job somewhere in CE or RB or in between so that "officialy" the job is not finished yet. It can even happen the the job will eventually fail in the last phases of submission (for example while uploading output sandbox), which does not prevent the other tests to succeed! Piotr On Sep 9, 2005, at 8:53 AM, Eric Fede wrote: > Hi > > Strange because i see the test job executed on the site. > > But i have a question because i don't understand something. > > When we see the results of the test (https://lcg-sft.cern.ch:9443/ > sft/sitehistory.cgi?site=marseillece01.mrs.grid.cnrs.fr), only the > job submission are on status ??, the other test are OK, or Warning > or ..... How is it possible to obtain the results of this tests > ( which run inside the site) if the job submission failed ? > > Eric > > > Frederic Schaer wrote: >> Hi, >> >> It seems that those sites which are overloaded (?) appear as "not >> tested" (the test job stays in queue forever, or stays running >> forever) because the test results are not returned. >> I submitted a job at 13H58, and it's still running (it has been so >> for hours now). >> >> Cheers, >> Fred >> >> Eric Fede wrote: >>> Hi Frederic >>> >>> The CPPM site is on JS error but when we have a look on the >>> history of the sft (https://lcg-sft.cern.ch:9443/sft/ >>> sitehistory.cgi?site=marseillece01.mrs.grid.cnrs.fr) ,we see that >>> the site is not monitored since yesterday evening. >>> >>> It is not the first time that we have this problem, the SFT stops >>> to test our site and continues to publish a old information. >>> >>> Eric >>> >>> >>> Frederic Schaer wrote: >>>> I see no difference on the 2 pages : >>>> >>>> lxb2001 : >>>> SD: 26 JL: 7 JS: 9 CT: 21 OK: 102 All: 165 >>>> >>>> lcg-sft : >>>> SD: 26 JL: 7 JS: 9 CT: 21 OK: 102 All: 165 >>>> >>>> ... Maybe there were differences because the results were beeing >>>> published (and loading the same page would have also given >>>> different results) ? >>>> >>>> Cheers, >>>> Fred >>>> >>>> Maarten Litmaath wrote: >>>>> Antun Balaz wrote: >>>>> >>>>>> Hi, >>>>>> I have just noticed that SFT results are not published anymore at >>>>>> >>>>>> https://lcg-sft.cern.ch:9443/sft/lastreport.cgi >>>>>> >>>>>> but rather at >>>>>> >>>>>> https://lxb2001.cern.ch:8443/sft/lastreport.cgi >>>>>> >>>>>> which was used previously. The first address was said to be >>>>>> permanent? >>>>> >>>>> In fact the results on the two pages are different: >>>>> the first has more sites and also a higher fraction of good >>>>> sites... >>>>> I suppose the SFT experts will comment on this. >>>>> >>> >>> >>> -- >>> -------------------------------------------------------------- >>> FEDE ERIC Mail : [log in to unmask] >>> CPPM Mail : [log in to unmask] >>> 163 Av de Luminy case 902 Tel : (+33) (0)4 91 82 76 41 >>> 13288 Marseille Cedex 9 France Fax : (+33) (0)4 91 82 72 99 >>> -------------------------------------------------------------- >>> > > > -- > -------------------------------------------------------------- > FEDE ERIC Mail : [log in to unmask] > CPPM Mail : [log in to unmask] > 163 Av de Luminy case 902 Tel : (+33) (0)4 91 82 76 41 > 13288 Marseille Cedex 9 France Fax : (+33) (0)4 91 82 72 99 > --------------------------------------------------------------