1549362 - Nightlies are not rendering any pages / IndexedDB errors

Status: VERIFIED FIXED

(Core :: Storage: localStorage & sessionStorage, defect)

Description

[Gian-Carlo Pascutto [:gcp]]

Current Nightly on Linux. When starting with an existing profile, no pages are loading.

Similar reports coming in from Windows, IndexedDB errors visible in console.

Comment 1

[Ryan VanderMeulen [:RyanVM]]

Doesn't reproduce with a clean profile for me, but does with a dirty one. Debug builds have a lot of warnings like this:

0:51.49 INFO: [Parent 9252, QuotaManager IO] WARNING: 'NS_FAILED(rv)', file z:/build/build/src/dom/localstorage/ActorsParent.cpp, line 380
0:51.49 INFO: [Parent 9252, QuotaManager IO] WARNING: 'NS_FAILED(rv)', file z:/build/build/src/dom/localstorage/ActorsParent.cpp, line 570
0:51.49 INFO: [Parent 9252, QuotaManager IO] WARNING: 'NS_FAILED(rv)', file z:/build/build/src/dom/localstorage/ActorsParent.cpp, line 7006
0:51.49 INFO: [Parent 9252, QuotaManager IO] WARNING: 'NS_FAILED(rv)', file z:/build/build/src/dom/localstorage/ActorsParent.cpp, line 7361
0:51.49 INFO: [Parent 9252, QuotaManager IO] WARNING: 'NS_FAILED(rv)', file z:/build/build/src/dom/localstorage/ActorsParent.cpp, line 6468
0:51.49 INFO: [Child 15584, Main Thread] WARNING: 'NS_FAILED(rv)', file z:/build/build/src/dom/localstorage/LSObject.cpp, line 823
0:51.49 INFO: [Child 15584, Main Thread] WARNING: 'NS_FAILED(rv)', file z:/build/build/src/dom/localstorage/LSObject.cpp, line 542
0:51.49 INFO: [Parent 9252, QuotaManager IO] WARNING: 'NS_FAILED(rv)', file z:/build/build/src/dom/localstorage/ActorsParent.cpp, line 380
0:51.50 INFO: [Parent 9252, QuotaManager IO] WARNING: 'NS_FAILED(rv)', file z:/build/build/src/dom/localstorage/ActorsParent.cpp, line 570
0:51.50 INFO: [Parent 9252, QuotaManager IO] WARNING: 'NS_FAILED(rv)', file z:/build/build/src/dom/localstorage/ActorsParent.cpp, line 7006
0:51.50 INFO: [Parent 9252, QuotaManager IO] WARNING: 'NS_FAILED(rv)', file z:/build/build/src/dom/localstorage/ActorsParent.cpp, line 7361
0:51.50 INFO: [Parent 9252, QuotaManager IO] WARNING: 'NS_FAILED(rv)', file z:/build/build/src/dom/localstorage/ActorsParent.cpp, line 6468
0:51.50 INFO: [Child 15584, Main Thread] WARNING: 'NS_FAILED(rv)', file z:/build/build/src/dom/localstorage/LSObject.cpp, line 823
0:51.51 INFO: [Child 15584, Main Thread] WARNING: 'NS_FAILED(rv)', file z:/build/build/src/dom/localstorage/LSObject.cpp, line 542

Comment 2

[Ryan VanderMeulen [:RyanVM]]

Backed out the below push so we can respin Nightlies:
https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?changeset=8dca7df2949204ed09622839bb0bbb0da11cbfca

https://hg.mozilla.org/mozilla-central/rev/df3eadfa74a8061f4c88496404d51baf47e21070

Comment 3

[Julien Cristau [:jcristau]]

Unsurprisingly things go back to working for me with dom.storage.next_gen=false.

Comment 4

[Andrew Overholt [:overholt]]

We should figure out which specific recent lsng change (the compression? the utf8 change?) caused this.

Comment 5

[Jan Varga [:janv]]

I know exactly what the problem is, but I had an appointment so I told sheriffs to back it out.
Working on a fix now.

Comment 6

[Jim Jeffery not reading bug-mail 1/2/11]

Now that the re-spin appears to be completed are updates still 'disabled' ? Updater says 'up to date'
Nevermind - just got the re-spin update.

Comment 7

[Ryan VanderMeulen [:RyanVM]]

Yes, respins are finished and updates are un-frozen. Users should be getting 20190506130308 now.

Comment 8

[Ryan VanderMeulen [:RyanVM]]

Jan is going to fix this in the regressing bug, so closing this as fixed by backout.

Comment 9

[Virtual_ManPL [:Virtual] 🇵🇱 - (please needinfo? me - so I will see your comment/reply/question/etc.)]

I'm confirming that bug is fixed, starting in latest Mozilla Firefox Nightly 68.0a1 (2019-05-06), so I'm marking this bug as VERIFIED.
Thank you very much for instant backout! \o/

Comment 10

[Firefox Bug Husbandry Bot]

Please specify a root cause for this bug. See :tmaity for more information.

Comment 11

[Jan Varga [:janv]]

It was an error in an SQL statement and we didn’t have a test for that.
Here are changes that fixed it: https://phabricator.services.mozilla.com/D29139?vs=99833&id=100039#toc