Page History
...
- recabling to srcf (cpo+ric+others)
- use data volume to each node to figure out cabling
- what should go through the bos? exclude opals?
- what should use MPO (Nov 1 checkpoint for laser locker/XPP; switch on Nov 16; revert by Nov 18 if unsuccessful) including BLD, IOC, FIM, XTCAV, Mono Enc TPR (we can switch daq timing independently of controls timing)
- IOC depends on Kukhee's software
- high rate BLD depends on Dawood's software
- Jeremy Mock writes about XTCAV: "Hi Chris, we can test the camera triggering soon, probably in October some time. However, I think XTCAV proper (the klystron) won't run on lcls2 timing until sometime next year. I think I heard maybe February...?". Pat Krejick is coordinating the change.
- 33kHz BLD: ebeam, gmd, test asap. waiting for software development (depends on Dawood)
- test reception of 33kHz gmd without beam ASAP working with Marcio
- commission s3df
- migrate ctl nodes (move hsd processes onto daq-tmo-hsd-01 (Ric), but not mono encoder)
- commission xpm0 in 208 (weaver)
- (would be nice) xtcav "nice" for acr? how do we share between tmo/rix: need bos-switchable timing?
- (would be nice) test psana2 live mode again (mona)
- training for scientists (e.g. on timing system and sequences) (matt)
- check we have psana2 detector interface for "sequence bits" (Matt thinks we have this: 272 bits)
- timing in all detectors
s3df to-do
- (done) psana1 (mikhail, valerio) psana2 (mikhail, mona, valerio)
- smalldata code (silke, vincent)
- will ask wilko to move some psana1 data over
- need path to install packages
- done by early february?
- movers (wilko)
- few days of work
- need some kafka work for movers and elog (wilko)
- should be ready for early february
- need some data management machines and irods integration (wilko)
- data storage (omar)?
- backend spinning-disk (aka "s3") 10's of PB done early january
- will get 1.5PB of flash free space by third of week of january
- serious issue: no extended ACL's since mounted via nfs. need multi-cluster feature? (expected in april?) or fully functional nfs?
- could workaround by making group the expt group, but then access via ps-data (e.g. silke) wouldn't work
- mon-node mount of calibdir (under /sdf/data/lcls) via weka-nfs (IT)
- will try third week of january
- ARP (murali, wilko)
- should be straightforward copy of what we currently have. confirm with murali.
- jupyter/ondemand (wilko, riccardo)
- could provide yee with default kernels (cpo will contact yee/wilko)
- can run in batch queues (charged) and interactive nodes (free!!!)
- (done) LCLS group permissions syncing (wilko)
- finalize high-priority analysis mechanism (preemption? reservations?)
- need coact (like iris at nersc) stuff from murali available by beginning of february
- will also coact to enable their s3df account
- enable user access to s3df (IT)
...
Overview
Content Tools