2024-02-01 Meeting Notes
Date
Feb 1, 2024
Participants
@Thomas, Jyothish (STFC,RAL,SC)
@Ian Johnson
@Alexander Rogovskiy
@Katy Ellis
@James Walder
@Thomas, Jyothish (STFC,RAL,SC)
Lancs: Gerard, Steven, @Matt Doidge
Glasgow:
Apologies:
CC:
Goals
List of Epics
New tickets
Consider new functionality / items
Detailed discussion of important topics
Site report activity
Discussion topics
Current status of Echo Gateways / WNs testing
Recent sandbox’s for review / deployments:
Item | Presenter | Notes |
|
---|---|---|---|
Operational Issues | @Thomas, Jyothish (STFC,RAL,SC) |
|
|
Gateways and WNs: | @Thomas, Jyothish (STFC,RAL,SC) |
Increase of memory to the proxy (XCache)to 64GB on the WN proxy due memory allocation errors. |
|
Rocky 8 migration planning | @Thomas, Jyothish (STFC,RAL,SC) |
To investigate missing rpms
|
|
S&C ATLAS week | @James Walder @Thomas, Jyothish (STFC,RAL,SC) | Next week S&C; Anything to discuss with Andy H et al. ? |
|
ECHO File transfer / throughput studies | @Katy Ellis | Updates?
|
|
Checksums fixes | @Alexander Rogovskiy | Status and plans for improving Checksumming work … GitHub - alex-rg/xrd_ckslib
|
|
Prefetch studies and WN changes | @Alexander Rogovskiy | Sandbox ready and applied to 1 WN, pending envroinment variable for timeout increase |
|
Deletion studies through RDR | @Ian Johnson | Latest deletion rates for 3 GiB files: 21 Hz Desire deletion rate from ATLAS: ~ 4 Hz Latest deletion rates for 1 KiB files: 59 Hz Deletion rates measured using 10 threads (VM with 16 vCPUs). In conversation with CERN DDM, Rucio reaper is using 10 threads (for ATLAS?). Next: Check timings for slow deletions. Measure deletion rate for 6 GiB files. Testing with 1 KiB files works up to 4000 files (deletion client on two VMs), but fails with 10000 files. Investigating. |
|
Tokens testing | @Thomas, Jyothish (STFC,RAL,SC) @Katy Ellis | https://stfc.atlassian.net/browse/XRD-63 avoid duplicating basepath by Jo-stfc · Pull Request #2151 · xrootd/xrootd |
|
Understanding CMSD Loadbalancing | @Thomas Byrne |
|
|
SKA Gateway box | @James Walder |
|
TEST | SVC16(R8)MEAN | SVC26(SL7)MEAN | SVC16(R8)MEDIAN | SVC26(SL7)MEDIAN |
---|---|---|---|---|
davsdelete | 0.47126 | 0.40214 | 0.464 | 0.381 |
davsread | 13.7352 | 12.7532 | 13.677 | 12.727 |
davsstat | 0.19786 | 0.18454 | 0.181 | 0.181 |
davswrite | 28.9813 | 23.9641 | 28.992 | 23.798 |
rootdelete | 0.27626 | 0.2847 | 0.258 | 0.253 |
rootread | 13.0728 | 12.0312 | 13.041 | 11.86 |
rootstat | 0.08458 | 0.07198 | 0.083 | 0.063 |
rootwrite | 25.3381 | 23.7827 | 25.312 | 23.738 |
tpcdest | 43.3166 | 40.3735 | 43.349 | 40.668 |
tpcsource | 37.468 | 24.8932 | 34.181 | 31.581 |
on GGUS:
Site reports
Lancaster - Tokens still disabled, but we reckon there will be pressure to re-enable. Pondering how to implement a network QoS to help reduce cephfs drop outs by guaranteeing a slither of bandwidth on the gateway hosts for ceph traffic. Also pondering dedicated internal gateways.
Glasgow - relatively stable, few network issues. OS/Ceph version update to do.
Action items
@James Walder to schedule a ‘hackathon’ within a F2F to have a session on architectural planning.
@James Walder to prepare an outline of the expected roadmap for XRootD developments in 2024.