Socket process enabled locally, but not on try server for android
Categories
(Core :: Networking, defect, P3)
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox75 | --- | fixed |
People
(Reporter: bdahl, Assigned: bwc)
References
(Blocks 1 open bug)
Details
(Whiteboard: [necko-triaged])
Attachments
(1 file)
|
(deleted),
text/x-phabricator-request
|
Details |
As mentioned here, when running mochitests locally the socket process is enabled for android. However, on try the socket process is disabled. This leads to discrepancies locally and on try and caused us to miss the linked bug.
Can we unify this setting so further problems aren't missed?
|
||
Updated•5 years ago
|
|
Assignee | |
Comment 1•5 years ago
|
||
Having a separate socket process variant for jobs on try ("spi") was a deliberate choice, since we wanted to ensure that both cases were tested. I guess any default we pick for local builds is not going to cover everything that runs in try.
|
Reporter | |
Comment 2•5 years ago
|
||
It doesn't look like "spi" is running on android.
|
||
Updated•5 years ago
|
|
|
Assignee | |
Comment 4•5 years ago
|
||
It looks like M-spi is also not running on linux64.
Edit: Actually, those jobs are for binary building only, but also run some browser-chrome mochitests? Weird.
|
|
Assignee | |
Comment 5•5 years ago
|
||
So I guess we need to modify this stuff:
It might break android testing that does not use e10s though. I can try once hg is back up again...
|
|
Assignee | |
Updated•5 years ago
|
|
|
Assignee | |
Comment 6•5 years ago
|
||
|
|
Assignee | |
Comment 7•5 years ago
|
||
|
|
Assignee | |
Comment 8•5 years ago
|
||
Um. The mda mochitests aren't running on android emulator at all. Since the mda tests are the only things that run on try with the socket process, even enabling the socketprocess variant on android does not result in M-spi jobs being scheduled...
|
|
Assignee | |
Comment 9•5 years ago
|
||
|
|
Assignee | |
Comment 10•5 years ago
|
||
|
|
Assignee | |
Comment 11•5 years ago
|
||
Try push looks ok, although there's an intermittent in a try push for bug 1613690. Looking into it.
|
|
Assignee | |
Comment 12•5 years ago
|
||
|
|
Assignee | |
Comment 13•5 years ago
|
||
Ok, have resolved the issue mentioned in comment 11. Try looks good now.
|
|
Assignee | |
Comment 14•5 years ago
|
||
Depends on D63198
|
|
Assignee | |
Comment 15•5 years ago
|
||
Hey, heads up -- we want to disable the socket server by default on Android (bug 1612961). You'll want to enable the network.process.enabled pref for the socketserver runs.
|
|
Assignee | |
Comment 17•5 years ago
|
||
(In reply to James Willcox (:snorp) (jwillcox@mozilla.com) (he/him) from comment #16)
Hey, heads up -- we want to disable the socket server by default on Android (bug 1612961). You'll want to enable the
network.process.enabledpref for the socketserver runs.
We go to great lengths to ensure that the socket process is enabled in CI if and only if we're running an spi job, so changes to defaults shouldn't be a problem for us I think.
|
||
Comment 18•5 years ago
|
||
|
||
Comment 19•5 years ago
|
||
| bugherder | ||
Description
•