836429 - Talos Regression tp4m_nochrome Main RSS, 3%, on Jan 29

Status: RESOLVED FIXED

(Firefox for Android Graveyard :: General, defect)

Description

[Geoff Brown [:gbrown]]

Reported in dev-tree-management Digest, Vol 49, Issue 119:

Message: 4
Date: Wed, 30 Jan 2013 17:21:52 -0000
From: [e-mail]
To: [e-mail]
Subject: <Regression> mobile - Tp4 Mobile NoChrome (Main RSS) -
        Android 2.2        (Native) - 2.98%
Message-ID:
        <[e-mail]>
Content-Type: text/plain; charset="us-ascii"

Regression: mobile - Tp4 Mobile NoChrome (Main RSS) - Android 2.2 (Native) - 2.98% increase
-------------------------------------------------------------------------------------------
    Previous: avg 93147350.000 stddev 736366.835 of 30 runs up to revision 6aaf13ffc716
    New     : avg 95921300.000 stddev 564365.316 of 5 runs since revision a31b07df79df
    Change  : +2773950.000 (2.98% / z=3.767)
    Graph   : http://mzl.la/YEC5YI

Changeset range: http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=6aaf13ffc716&tochange=a31b07df79df

Comment 1

[Geoff Brown [:gbrown]]

Looking back at the m-i graph, it seems this measure has been slowly increasing for several months, up from about 86M at the beginning of November 2012 to about 96M today.

http://graphs.mozilla.org/graph.html#tests=[[136,63,20]]&sel=none&displayrange=90&datatype=running

Comment 2

[Geoff Brown [:gbrown]]

This particular regression seems to happen on m-i at:

Rev e9a03227ae72
Jan 29, 2013 16:43
92784300

Rev 009b00bcce30
Jan 29, 2013 17:10
96067000
Δ 3282700 (3.5%)


On m-c graphs, the change is not as apparent, but I note:

Rev 6aaf13ffc716
Jan 29, 2013 20:30
92648500

Rev a31b07df79df
95959000
Δ 3310500 (3.6%)
Jan 30, 2013 03:13

Comment 3

[Geoff Brown [:gbrown]]

Caused by bug 761552?

Comment 4

[Kartikaya Gupta (email:[e-mail])]

I backed out the offending change in https://hg.mozilla.org/integration/mozilla-inbound/rev/8728de36d4a8. Initial tpn results on inbound seem to confirm that the memory usage has dropped back down.