2010/11/11 Piotr Drąg <piotrdrag(a)gmail.com>:
2010/11/10 Domingo Becker <domingobecker(a)gmail.com>:
> It seems that a script to refresh/clean the module's cache is not
> working from cron for some reason.
> I've been talking about this issue to diegobz at #fedora-latam-rpmdev
> and he said we should ping fedora-admin guys at the appropriate
> If they can't solve the issue, they will ask upstream (diegobz and
> glezos), and they will help solving it.
> I think we need to have a wiki page to explain translators how to ask
> for help to fedora-admin people about Transifex issues. Perhaps it
> exists, but nobody seems to find it.
It has been filed several times at different places (Infra trac,
Bugzilla, mailing lists), but sure we can do it once more. I'll file a
new ticket about cron job at Infra trac instance tomorrow after work,
if no one will do it to then.
I mean it should be easy for a translator with this problem to find
out how to get help from admin people.
If a document already exists commenting how to do it, then it is not
reached by translators facing this problem, and that's why they are
using this list. Admin people normally do not read this list, or if
they do, they expect us to follow their rules.
What can we do to improve this?
I guess a reference at the L10N page  to another page that document
how to proceed in case of a Transifex issue.
If that page already exists, it's just a matter to add a link.
The fact that this error was informed before and it was solved by
someone, doesn't mean that it will not appear anymore.
It would probably be fixed by an upgrade. It's what we are trying to do.
In the meantime, in order to keep the current Transifex instance
alive, errors should be informed as many times as necessary.
It think the real problem is (and I may be wrong) that the cron job
for refreshing cache is overwhelmed by the number of modules,
languages and files it actually has.
If the cron job runs four times a day, with a very low priority to let
translators work in the meantime, it's probably not enough time to
It may start over and over again, resulting in a deadlock and
consuming too much resources.
If I see a cron job that starts over and over again without
terminating the previous one, I would disable it, no matter what it
does. Perhaps that's what it's happening.
But, as I said, I may be wrong.
Domingo Becker (es)