TrigMC software release coordination: PTD's logbook - page for September 2018

Notes

Use the shift date in reverse chronological order to enter any new Merge Requests.

For each shift date show the NEW MRs, and any related relevant comments/questions; note down when I (or some one else) merged/closed each request, and mark the MR as done/closed by ticking the box ([x]). If an MR takes longer than a day to merge, leave it listed under the original shift date, but add comments prefaced with the date of the comment/action. Only MRs that are still active should have the box unticked ([ ]).

Therefore each MR will only appear once in the log, under the date of the day when I first became aware of it.

27/9/2018

[x] !14523 Update 21.0-TrigMC to 21.1.41 (opened by Julie Kirk)

2.00pm Request opened yesterday, but pipeline failed; some changes were made, and pipeline is running again (started 30 mins ago).

7.15pm Pipeline has now succeeded. However Julie asked Rui Zou (on the MR page) to OK her changes, but he hasn't replied yet. Will ask them if the issue is OK, or should I wait for confirmation before merging...

7.41pm Julie replied that, even though she has not got an answer from Rui yet, she would rather we merge now. If it turns out her changes should not have been made, she can undo them (it just boils down to putting back two trigger chains that she removed) at the next TrigMC MR. Merged OK.

20/9/2018

[x] New release out of 21.0-TrigMC 2018-09-19T2138: release/21.5.8

11.23am Checked the nightly, in advance of a possible request for a build today. Noticed that in addition to last night's nightly, the only previous nightly in the system is for 13/9 (so it hasn't been building nightly...?). Last night's nightly has 3 Athena ART errors/warnings, but they are exactly the same as for the 13/9 nightly, so nothing new there.

11.26am Correction: upon looking more carefully (here: http://atlas-computing.web.cern.ch/atlas-computing/links/distDirectory/gitwww/GITWebArea/.testp/21.0-TrigMC/index.html). I can see that the TrigMC nightlies have indeed been built every night for a week... For some reason, in the main screen (http://atlas-computing.web.cern.ch/atlas-computing/links/distDirectory/gitwww/GITWebArea/.testp/) it only shows two (19/9 and 13/9). The remarks below still stand: same errors/warnings in most recent nightly as in the previous night's one.

4.50pm emailed Rafal Bielski to ask if he is still interested in a new release being made out of last night's nightly, or if he needs more time to check the results, before deciding on this.

5.11pm Rafal has replied that they are waiting for Ben Carlson to verify that the change that he committed a couple of days ago (see MR !14295, below) is working as intended. Once that is established Ben will give the greenlight to go ahead with the release.

21/9/2019 10.01am Ben Carlson and Rafal have confirmed that they wish to have a new TrigMC release. This will be release 21.5.8, off the 21.0-TrigMC 2018-09-19T2138 nightly build.

21/9/2018 10.27am Started the job on Jenkins to copy the RPM files over to the build location.

21/9/2018 10.32am Opened JIRA ticket with the request; ticket assigned to Alessandro de Salvo.

21/9/2018 10.47am New release tagged, and release candidate number updated. The latter will be effected via an MR (!14406) assigned to me as usual. Pipeline started running immediately; merged it immediately, to avoid wasting resource on running pipeline for a trivial change. This caused the pipeline to "fail" and stop, as intended.

21/9/2018 4.10pm New release has been installed in the official location, and also deployed on the Grid. Have just now emailed the developers and Trigger release coordinators to let them know.

[x] !14347 Sweeping 14285 from 21.1 to 21.0-TrigMC. Update TrigAnalysisTest tests: "trigedm_grid" and "trigdecisiontool_grid" (opened by Atlas Nightlybuild)

11.12am Another automatic sweep from 21.1. Pipeline OK, and already previously approved anyway, so will merge now. Merged OK.

[x] !14342 Sweeping 14332 from 21.1 to 21.0-TrigMC. Remove TrigNavSlimming post processing for "rdotoesdaod" ART tests in AthenaP1 branches. Fixes ATR-18611 (opened by Atlas Nightlybuild)

11.09am This sweep from 21.1 was opened automatically overnight. CI pipeline has succeeded. As it is a sweep and the original MR has previoulsy been approved, will merge now. Merged OK.

19/9/2018

[x] !14314 Sweeping 14061 from 21.0 to 21.0-TrigMC. transform+job options for Tower22 Slice A data (opened by Atlas Nightlybuild)

11.10am Another MR opened overnight automatically, as MR !14061 was tagged "alsoTargeting:21.0-TrigMC". Pipeline started 4 hrs ago; still running.

11.45am Pipeline failed: "make", due to an error creating a symbolic link ("file already exists").

20/9/2018 11.45am Peter Berta has looked into this issue and is of the opinion that the error that caused the pipeline fail is not specific to the MR, and therefore will have another go at running the pipeline (started running 3 hrs ago.)

20/09/2018 3.50pm CI pipeline has now succeeded. Merged OK.

[x] !14312 Sweeping 13983 from 21.0 to 21.0-TrigMC. Step-by-step efficiency (opened by Atlasnightlybuild)

11.01am Opened overnight by Atlas Nightlybuild (about 8hrs ago), as MR !13983 to 21.0 had an "alsoTargeting:21.0-TrigMC" label. Note that, correspondingly, the present MR is tagged "sweep:from 21.0". Pipeline was started "by hand", only about 2 hrs ago and is still running.

4.06pm Pipeline succeded 45 mins ago. Merged OK.

18/9/2018

[x] !14295 allow for masking good ROBS (opened by Ben Carlson)

7.15pm Request opened 2 hrs ago. Pipeline has been running for one hour. MR won't make it into tonight's build. Rafal Bielski appended a note indicating that once/if the MR goes through and following nightly tests are OK, then they will be requesting a new build (21.5.8)... If all goes OK, it should be sometime on Thursday.

11.56pm CI pipeline passed. Merged OK.

17/9/2018

[x] !14270 Merge release/21.0.80 into the branch 21.0-TrigMC (opened by Rafal Bielski)

9.20pm Opened about 45 mins ago. Pipeline is running.

18/9/2018 9.32am CI pipeline passed. [John Chapman added the sweep:ignore label an hour ago.] Merged OK.

13/9/2018

[x] !14164 Sweeping 14131 from master to 21.0-TrigMC. TrigFTKSim: Avoid cling warnings. (opened by Atlas Nightlybuild)

12.42pm MR opened overnight. Changes no approved by L1 shifter. Merged OK.

11/9/2018

John Chapman asked about the frequent "sweep (to master) failed" messages. See discussion under MR 13995.

8/9/2018

[x] !14064 Sweeping 14043 from 21.1 to 21.0-TrigMC. Use default input (pp) in ART tests for HI_v4 menu (opened by Atlas Nightlybuild)

11.19am MR opened overnight. CI pipeline has run, and passed a few hours ago. Merged OK (...from the train in Portugal, Lisbon - Coimbra...!)

5/9/2018

[x] !13995 Update 21.0-TrigMC to 21.0.79 (opened by Rafal Bielski)

10.41am MR opened 10 mins ago. CI pipeline started.

6.03pm CI pipeline has concluded successfully. Accepting MR, as all changes have been approved before anyway. Merged OK.

-- PedroTeixeiraDias - 02 Sep 2018