On 7 Sep 2010, at 09:36, Hellier, Richard (STFC,RAL,ESC) wrote:
> At RAL we’ve been seeing strange problems with our top-level BDII servers. Chris Walker at QMUL logged a GGUS ticket about some of Steve Lloyd’s tests failing (on doing replication for example) because required entries were absent from the information service.
Just in case you hadn't noticed; repeated queries to that BDII get different information.
Specifically, it looks like some of the front end nodes [0] have the data for Glasgow, and some don't.
It's not a critical problem; but it is causing a lot of false alarms on the ops dashboard. I can stomp on them, of course, but I thought I should let some one know (and I hope you're the right person...., in case it's a symptom of a deeper problem than a simple replication failure.
Stuart Purdie
ROD and UKI-SCOTGRID-GLASGOW site admin
[0] I'm thinking that the address is loadb alanced to a number of servers somehow - but I can't recall the details, so I might be wrong on that.
|