I'm seeing lots of tests failing with "too many connection resets (due to Net::ReadTimeout - Net::ReadTimeout) after 2 requests on 70020846470700, last used 60.008243583 seconds ago (Net::HTTP::Persistent::Error)"
e.g. https://wmf.ci.cloudbees.com/job/MobileFrontend-en.m.wikipedia.beta.wmflabs....
Long term view: How do we stop these? Short term view: It would be useful if no email notifications were generated when this happens or these failures were filtered from test results...
On Tue, Mar 25, 2014 at 3:34 PM, Jon Robson jdlrobson@gmail.com wrote:
I'm seeing lots of tests failing with "too many connection resets (due to Net::ReadTimeout - Net::ReadTimeout) a Long term view: How do we stop these?
We port our Jenkins builds for browser tests from Cloudbees to WMF Jenkins. This is in progress now. We have a few issues to sort along the way.
The problem is that we don't have any information about the connection from Jenkins on Cloudbees to Sauce Labs hosts. Željko and I have filed a number of support tickets with Cloudbees and we have made no progress.
When we run our own builds we will have all of the information about our end of that connection, when today we have no information about either end of that connection.
-Chris
Okay thanks Chris - is there a bug we can track this?
On Tue, Mar 25, 2014 at 3:47 PM, Chris McMahon cmcmahon@wikimedia.org wrote:
On Tue, Mar 25, 2014 at 3:34 PM, Jon Robson jdlrobson@gmail.com wrote:
I'm seeing lots of tests failing with "too many connection resets (due to Net::ReadTimeout - Net::ReadTimeout) a Long term view: How do we stop these?
We port our Jenkins builds for browser tests from Cloudbees to WMF Jenkins. This is in progress now. We have a few issues to sort along the way.
The problem is that we don't have any information about the connection from Jenkins on Cloudbees to Sauce Labs hosts. Željko and I have filed a number of support tickets with Cloudbees and we have made no progress.
When we run our own builds we will have all of the information about our end of that connection, when today we have no information about either end of that connection.
-Chris
Mobile-l mailing list Mobile-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/mobile-l
On Tue, Mar 25, 2014 at 11:54 PM, Jon Robson jrobson@wikimedia.org wrote:
Okay thanks Chris - is there a bug we can track this?
Maybe this one:
https://bugzilla.wikimedia.org/show_bug.cgi?id=60338
Željko
I actually am talking to Cloudbees about this right now, but I don't have high hopes for a resolution. -C
On Thu, Mar 27, 2014 at 8:41 AM, Željko Filipin zfilipin@wikimedia.orgwrote:
On Tue, Mar 25, 2014 at 11:54 PM, Jon Robson jrobson@wikimedia.orgwrote:
Okay thanks Chris - is there a bug we can track this?
Maybe this one:
https://bugzilla.wikimedia.org/show_bug.cgi?id=60338
Željko
On Tue, Mar 25, 2014 at 11:34 PM, Jon Robson jdlrobson@gmail.com wrote:
I'm seeing lots of tests failing with "too many connection resets (due to Net::ReadTimeout - Net::ReadTimeout) after 2 requests on 70020846470700, last used 60.008243583 seconds ago (Net::HTTP::Persistent::Error)"
I think this is caused by this net-http-persistent bug[1], but unfortunately I have never had the time to investigate the problem in detail.
Željko -- 1: https://github.com/drbrain/net-http-persistent/issues/37