Re: REST-API: Merging/Rebasing with Mercurial?

On Wed, Jan 30, 2019 at 08:53:40PM +0000, Joseph Myers wrote:
> The Travis CI tests appeared to have failed, although I think that's 
> unrelated to these changes.
> https://travis-ci.org/roundup-tracker/roundup/builds/486608465

I first thought these are the GPG tests, see my other message to the
list and the Results Berhard had reproducing these failures.
Seems travis is *not* failing in GPG and is failing some mysql tests
instead. So I may have incompatible gpg-related packages installed.

The failing mysql tests all have an "Operational Error 2006" which seems
to mean "Database has gone away". They happen all during database
tearDown, in a state where all asserts of the tests were already
successful. But if there is a problem with the db connection I'd expect
this not to be reproduceable in exactly the same test-cases.

The first google hit for that error yields
"The MySQL server has gone away (error 2006) has two main causes and
solutions: Server timed out and closed the connection. To fix, check
that wait_timeout mysql variable in your my.cnf configuration file is
large enough. You may also need to increase the innodb_log_file_size
mysql variable in your my.cnf configuration to for example 128MB or
higher." There is a third reason mentioned that I don't think applies in
our case.
https://matomo.org/faq/troubleshooting/faq_183/

So maybe we have some timing variations on the travis server where
sometimes mysql loses patience and closes the connection?
Or these tests produce large results that tax the mentioned log file
size limit?

Ralf
-- 
Dr. Ralf Schlatterbeck                  Tel:   +43/2243/26465-16
Open Source Consulting                  www:   http://www.runtux.com
Reichergasse 131, A-3411 Weidling       email: [email protected]