Difference: ComputingT2Transfers (8 vs. 9)
Revision 92014-09-12 - samir
| Line: 1 to 1 | ||||||||
|---|---|---|---|---|---|---|---|---|
USCMS T2 Transfers | ||||||||
| Line: 14 to 14 | ||||||||
Spreadsheet | ||||||||
| Added: | ||||||||
| > > | As of 12th of September, the only site which managed interesting rates was Purdue. Most SE to SE links won't pass ~70 transfers and 500~1000 MBps, which is a bit disappointing. Suggesting to the group that we try a general FTS-optimizer bypass if everyone is comfortable with throttling transfers in their download agents. This is important as otherwise the number of transfers can grow out of control, crashing either GridFTPs (site SE) or the PhEDEx host managing those. | |||||||
Site notes | ||||||||
| Added: | ||||||||
| > > | CaltechStarted having consistent stage-out problems by Sep 11th. Throttled down the transfers to other sites so we can investigate the issue with less load on the GridFTPs. | |||||||
NebraskaHad good rates in general, but didn't see more than 1000 MBps from Caltech.Purdue | ||||||||
| Changed: | ||||||||
| < < | Found that there was a problem related to LHCONE peering. Their traffic from Caltech was going through CENIC, not LHCONE, which is sub-optimal. Purdue is following up with local network support. | |||||||
| > > | By Sep 11th the site started doing very interesting download rates. Up to 28 Gbps download rates averaged over 10 min.
Plots here
= Older notes = Found that there was a problem related to LHCONE peering. Their traffic from Caltech was going through CENIC, not LHCONE, which is sub-optimal. Purdue is following up with local network support. | |||||||
| Even though there was this problem, we still observed decent rates > 500 MBps at some distinct moments. | ||||||||
View topic | History: r9 < r8 < r7 < r6 | More topic actions...
Copyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors. Ideas, requests, problems regarding TWiki? Send feedback