692346 - explicitly require a /mnt/sdcard-based tegra devRoot

Status: RESOLVED FIXED

(Release Engineering :: General, defect)

Description

[Aki Sasaki (not active)]

tegra-033 seems to not have an sdcard, so it's giving a bogus devRoot:


Connecting to: 10.250.49.20
reconnecting socket
send cmd: testroot

recv'ing...
response: /data/data/com.mozilla.SUTAgentAndroid/files


This is causing a ton of oranges.

Right now we're fine as long as devRoot is not None and not /tests.  We need to barf and set error.flg if not devRoot.startswith("/mnt/sdcard").


Taking tegra-033 out of production...

Comment 1

[Joel Maher ( :jmaher ) (UTC -8)]

yes, this has been seen as the cause of half of the oranges I looked at on tbpl tonight.  Not saying that if this was fixed the test would be green:)

Comment 2

[Joel Maher ( :jmaher ) (UTC -8)]

from logcat:
D/Vold    (  924): Volume sdcard state changing -1 (Initializing) -> 0 (No-Media)

Comment 3

[Aki Sasaki (not active)]

Attachment: devRoot must startwith("/mnt/sdcard")

The first instance of this will go purple, which isn't great.
However, now it will set the error flag and disable the tegra from running any more jobs, which reduces the oranges quite a bit.

Not sure if we want startswith("/mnt/sdcard") since that'll match other longer strings, but I don't see it being a glaring problem either.

Comment 4

[Joel Maher ( :jmaher ) (UTC -8)]

Comment on attachment 565110 [details] [diff] [review]
devRoot must startwith("/mnt/sdcard")

Review of attachment 565110 [details] [diff] [review]:
-----------------------------------------------------------------

nice and simple!

Comment 5

[Aki Sasaki (not active)]

Comment on attachment 565110 [details] [diff] [review]
devRoot must startwith("/mnt/sdcard")

http://hg.mozilla.org/build/tools/rev/0bd747e8330a

I'll push these out to the foopies.

Comment 6

[Aki Sasaki (not active)]

Updated /builds/tools on foopy06 - 14.