Closed Bug 349183 Opened 18 years ago Closed 18 years ago

Update outage - reflow tinderbox published things it shouldn't have

Categories

(AUS Graveyard :: Systems, defect)

Product:
AUS Graveyard
Type:
defect
Priority:
Not set
Severity:
major

Tracking

(Not tracked)

Status:
RESOLVED FIXED
Milestone:
4.x (triaged)

People

(Reporter: nthomas, Assigned: rhelmer)

Details

The new Reflow tinderbox is pushing update info into the nightly channel for the Firefox trunk, which looks to have busted updates across the board. Partial generation for all the apps & branches is failing, presumably because patcher can't find the builds. Firefox trunk testers also got updated from the 2006081704 nightly to the -1709 reflow build, and were surprised to find that Firefox now passed Acid2. :-) Props to them for figuring out what happened. As far as I can tell, the steps to fix this up are: 1, On fxexp-win32-tbox : Fx-Reflow, unset $update_pushinfo [Rob Helmer has done this already] 2, In the AUS datastore, remove the reflow builds Firefox/trunk/WINNT_x86-msvc/2006081709/* Firefox/trunk/WINNT_x86-msvc/2006081716/* Firefox/trunk/WINNT_x86-msvc/2006081720/* Firefox/trunk/WINNT_x86-msvc/2006081804/* The geniune nightly from gaius will be an -1805 buildtime (based on the start of the build). 3, Make Firefox/Trunk/Windows update properly. I'm guessing here, but probably zap on AUS Firefox/trunk/WINNT_x86-msvc/2006081704/en-US/partial.txt and remove from stage.m.o pub/mozilla.org/firefox/nightly/2006-08-17-04-trunk/firefox-3.0a1.en-US.win32.partial.2006081704-2006081709.mar
(In reply to comment #0) > The new Reflow tinderbox is pushing update info into the nightly > channel for the Firefox trunk, which looks to have busted updates > across the board. Partial generation for all the apps & branches is > failing, presumably because patcher can't find the builds. Firefox > trunk testers also got updated from the 2006081704 nightly to the > -1709 reflow build, and were surprised to find that Firefox now passed > Acid2. :-) Props to them for figuring out what happened. > > As far as I can tell, the steps to fix this up are: > 1, On fxexp-win32-tbox : Fx-Reflow, unset $update_pushinfo > [Rob Helmer has done this already] > > 2, In the AUS datastore, remove the reflow builds > Firefox/trunk/WINNT_x86-msvc/2006081709/* > Firefox/trunk/WINNT_x86-msvc/2006081716/* > Firefox/trunk/WINNT_x86-msvc/2006081720/* > Firefox/trunk/WINNT_x86-msvc/2006081804/* This is done now as well. > The geniune nightly from gaius will be an -1805 buildtime (based on > the start of the build). > > 3, Make Firefox/Trunk/Windows update properly. I'm guessing here, but > probably zap on AUS > Firefox/trunk/WINNT_x86-msvc/2006081704/en-US/partial.txt > and remove from stage.m.o > pub/mozilla.org/firefox/nightly/2006-08-17-04-trunk/firefox-3.0a1.en-US.win32.partial.2006081704-2006081709.mar > I removed the snippet, I don't think it's necessary to remove the MAR file. I set gaius to do clobber (which pushed updates) next cycle, so trunk users should update back to the latest normal (non-reflow-refactor-branch :) ) build.
(In reply to comment #0) > 2, In the AUS datastore, remove the reflow builds > Firefox/trunk/WINNT_x86-msvc/2006081709/* > Firefox/trunk/WINNT_x86-msvc/2006081716/* > Firefox/trunk/WINNT_x86-msvc/2006081720/* > Firefox/trunk/WINNT_x86-msvc/2006081804/* > The geniune nightly from gaius will be an -1805 buildtime (based on > the start of the build). Nick pointed out to me in IRC that the nightly update generator was failing here, this needed to be removed on there as well.
This seems fixed now; should it be marked?
All builds from the trunk and Mozilla1.8 branch are updating ok now, so we can close this.
Status: NEW → RESOLVED
Closed: 18 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.