From: Jerome L. <ad...@gm...> - 2014-08-12 00:59:18
|
Hi Friedel This issue looks completely unrelated to nginx. Investigating. J. Leclanche On Mon, Aug 11, 2014 at 7:31 PM, F Wolff <fr...@tr...> wrote: > On Mon, Aug 4, 2014 at 4:35 PM, Jerome Leclanche <ad...@gm...> wrote: >> Hi Friedel >> >> Amagama was restarted the other day. I wanted to recap the situation though. >> >> There was a potential security issue in Amagama's server-side input >> validation that would fail a transaction with the database. The >> transaction was never being rolled back and amagama would enter a >> permanently broken state. > > Thank you for explaining, Jerome. > > >> This was "hidden" in apache because apache restarts mod_wsgi now and >> then (brought to you by the people who reboot their computer instead >> of fixing memory leaks). When we switched to nginx, we stopped doing >> that nonsense. That meant this bug would be more visible. >> It has now been properly fixed. It's likely that amagama's previous >> erratic behaviour was due to it. > > I guess it was "hidden" in the error logs, but I guess nobody had an > opportunity to review it, and it seemed to be running fine. > > Amagama is again down now, and I guess it is because it doesn't > automatically recover from bad situations. It seems that this new > setup is very fragile, and additionally needs a lot of hand-holding. > As far as I know it was running fine for years under the previous > setup. I'm still not sure what the major advantage is of this way. > > Anyway, can someone with the necessary powers and understanding > restart the server, please? > > Keep well > Friedel > > ------------------------------------------------------------------------------ > _______________________________________________ > Translate-devel mailing list > Tra...@li... > https://lists.sourceforge.net/lists/listinfo/translate-devel |