1214595 - IndexedDB actorsParent crash on Z3c after x-heavy reference workload

Status: RESOLVED WORKSFORME

(Core :: Storage: IndexedDB, defect)

Description

[:gerard-majax]

Attachment: gdb stacktrace

STR:
 0. Build & flash Z3c L opt
 1. make reference-workload-x-heavy

Expected:
 Boots properly

Actual:
 Segfault

Comment 1

[:gerard-majax]

Attachment: gdb session content

Comment 2

[:gerard-majax]

As far as I can say, it is 100% repro when starting from a clean state.

Comment 3

[:gerard-majax]

We might be exactly hitting what Kyle said on bug 1203803 comment 123

Comment 4

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

I don't think this is related to bug 1203803.

Comment 5

[Ken Chang[:kenkai]]

It is for L. Should it be a blocker?

Comment 6

[:gerard-majax]

(In reply to Ken Chang[:ken] from comment #5)
> It is for L. Should it be a blocker?

I had the issue on L because I was working on something else, but I doubt it's because of L specifically ...

Comment 7

[Gregor Wagner [:gwagner]]

I reproduced on kk as well.

Comment 8

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

The error code nsresult here is NS_ERROR_DOM_INDEXEDDB_ABORT_ERR.

Comment 9

[Andrew Overholt [:overholt]]

Jan, can you look into this?  If the workload reflects real life usage, it seems like we should block B2G 2.5 on it.  Jan, that means we have ~1.5 weeks :/

Comment 10

[Andrew Overholt [:overholt]]

Jan is now set up with a b2g environment and is working on this.

Comment 11

[Mahendra Potharaju [:mahe]]

Jan, Can you please post any progress on this bug? Thanks

Comment 12

[Jan Varga [:janv]]

I can't reproduce it anymore. Bug 961049 or something else might have fixed it.
Can someone verify it ? Thanks.

Comment 13

[Mahendra Potharaju [:mahe]]

Adding NI to Alexendre/Gregor to try reproducing it.

Comment 14

[:gerard-majax]

Looks like I don't reproduce anymore.

Comment 15

[Jan Varga [:janv]]

(In reply to Alexandre LISSY :gerard-majax from comment #14)
> Looks like I don't reproduce anymore.

Wow, thanks for verifying!

Comment 16

[Andrew Overholt [:overholt]]

If bug 961049 fixed this (a 2.5+ bug), do we need to uplift bug 961049 to 2.5? It's a big change ...

Comment 17

[:gerard-majax]

(In reply to Andrew Overholt [:overholt] from comment #16)
> If bug 961049 fixed this (a 2.5+ bug), do we need to uplift bug 961049 to
> 2.5? It's a big change ...

So I took gecko at the commit just before bug 961049 landed, and I don't repro the issue either.

Comment 18

[Jan Varga [:janv]]

(In reply to Alexandre LISSY :gerard-majax from comment #17)
> (In reply to Andrew Overholt [:overholt] from comment #16)
> > If bug 961049 fixed this (a 2.5+ bug), do we need to uplift bug 961049 to
> > 2.5? It's a big change ...
> 
> So I took gecko at the commit just before bug 961049 landed, and I don't
> repro the issue either.

Ok, so something else must have changed. Hm...

Comment 19

[Mahendra Potharaju [:mahe]]

Based on comment 17, we will not uplift bug 961049 to 2.5