 |
 |
Hi Steve, Our emails just passed! Anyway, quite likely a bug. I can escalate concerns about the reporting, but once we’ve uncovered the issue you’ll have to trigger the new computation. Cheers, Jeremy > On 7 Dec 2016, at 22:15, sjones <[log in to unmask]> wrote: > > Ah.. I can see John's already replied. > > It could be some (obscure perhaps?) monitor check that was not reported to us via a ticket. Suggest either: > a) monitors that detect problems should send alerts/tickets or > b) sites to be informed of any monitor used to make rel/av tests which need to be manually observed on a regular basis. > > In any case, we were up and running - the estimates are likely to be false ones. > > Cheers, > > Ste > > > On 2016-12-07 22:02, sjones wrote: >> Hi Jeremy, >> On 2016-12-07 16:52, Jeremy Coles wrote: >>> Please could: RHUL, Glasgow and Liverpool send me some brief text on >>> the difficulties encountered during the month. >> There were no significant difficulties in November at Liverpool >> regarding availability and reliability prior to a power cut in late >> November that wiped out our ARC/Condor CE, but we recovered within 6 >> hours or so. Thus we did practically a full month of work at near 100% >> reliability & availability. >> My figures suggest we made 6.3 million hs06 hours of work for lhcb in >> Nov, which is indicative of (say) 99% uptime. It would not be possible >> if we were up only 33% of the time. I'll check this in the morning but >> this looks like a serious measurement error to me. >> To do that, I need to know; what is the basis of the 33% measurement, >> and who is responsible for making it? >> Cheers, >> Ste