[Webkit-unassigned] [Bug 33153] Running multiple instances of run-webkit-tests is not possible on the same machine

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Mon Jan 4 13:42:46 PST 2010


https://bugs.webkit.org/show_bug.cgi?id=33153





--- Comment #10 from Andras Becsi <abecsi at inf.u-szeged.hu>  2010-01-04 13:42:43 PST ---
(In reply to comment #7)
> Supporting parallelism inside run-webkit-tests, and also being able to run
> several instances of the script sounds like overkill. Did you consider running
> multiple buildbots in virtual machines?
We use virtual machines now, but using them means significant loss of
horsepower and is a huge wastage of energy, time and the capabilities of our
server, so, on the contrary: using virtual machines only for running the tests
is overkill, as I see it.

> In any case, making http and websocket tests port-agnostic is not practical.
> Most of them are not, and those that are got significantly more complicated
> because of that. Putting the burden of writing port-agnostic tests on
> developers is a cost we'd have to keep paying in the future, and thus
> undesirable.
I do not think that changing related tests to use and expect a preconfigured
port number (which would be fixed per-buildbot in our usecase) would make
writing tests significantly complicated.

> There are more difficulties with running multiple instances of the script than
> mentioned in this bug - please see bug 10906 for discussion.
On our development servers we are running multiple instances of
run-webkit-tests --no-http parallel without any problems except of very rare
issues related to the /tmp/LayoutTests symlink.

Bug 10906 seems quiet dead and is dealing with a much more complicated issue,
which wouldn't help us to maximalize the efficiency of our hardware, and help
other developers who want to test on their shared developer hardware resource
more effectively.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.



More information about the webkit-unassigned mailing list