On 18/09/12 09:56, Christopher J.Walker wrote: > On 18/09/12 09:44, Mark Slater wrote: >> Apologies for the noise - that should have just gone to Jeremy :) >> >> Mark >> >> On 18/09/12 09:43, Mark Slater wrote: >>> Hi Jeremy, >>> >>> I'm afraid I won't be able to join the meeting as I'm helping to run >>> the PPAP conference in Bham at present! THere's not much to report >>> from LHCb other than reprocessing is under way and progressing. >>> Sunday/Monday we had quite a few failed pilots at RAL for which I've >>> filed a ticket: >>> >>> https://ggus.eu/ws/ticket_info.php?ticket=86156 >>> >>> >>> Tiju says it's under investigation. I haven't had chance to go over >>> the other UK sites but I see there's is some movement on some tickets >>> (though I understand there's some issue with a single ticket being >>> used for different problems!) >>> >>> For Bham, I'd be interested to know if there's any discussion about >>> both ours and Oxfords poor copying rate to BNL. We can only think it's >>> something to do with JANET! >>> >>> https://ggus.eu/ws/ticket_info.php?ticket=86105 >>> > > Can you (and Oxford) do traceroutes to and from BNL's perfsonar please > (from your cluster). It might be useful to post them here - as a follow > up to my routes request. > > QMUL is seeing low transfer rates from Taiwan - fell from 5MB/s to 0.7 > when we got the 10Gig link and we believe it's due to traffic to our > cluster taking a suboptimal route. > https://ggus.eu/ws/ticket_info.php?ticket=86107 has details. > The other piece of generic advice is: http://fasterdata.es.net/ Can you confirm you've applied the recommended TCP settings there. Chris