________________________________
From: LHC Computer Grid - Rollout on behalf of Rod Walker
Sent: Wed 5/4/2005 10:43 PM
To: [log in to unmask]
Subject: Re: [LCG-ROLLOUT] [Fwd: FW: RLS information]
Hi,
Well pass on our thanks for all the RLS stabilization work. We still need
to understand where this extra load is coming from. The production
activities have wound down if any change at all. Recently data replication
using dq reliable file transfer have started, but already for a week or
so. The number of concurrent transfers is/was limited to 20, so I
can`t see that this is the cause.
My favourite would be that some punters are looking for Rome data, using
some script to cycle through the datasets. Is there any indication which
node the load is coming from (at least from netstat).
Does anyone know which component is doing COUNT (*), i.e. counting the
number of replicas?
I get communication errors right now, so a netstat on the rls node would
be useful.
Cheers,
Rod.
On Wed, 4 May 2005, Maarten Litmaath wrote:
--
Rod Walker +1 6042913051
|