Re: pl1.php.net Disabled
[email protected] (Hannes Magnusson)
| Newsgroups | php.mirrors |
|---|---|
| Message-ID | <CADNQb0X2hXGAsNuHyK0iNAF2=Z=Qoe7YTcn=9JnjkGL9vTbTbg@mail.gmail.com> |
Mirrors that are frequently randomly unavailable are very frustrating for users. If our mirror check system observes down time and sends you a notification -- imagine the annoyance from a real user not being able to reach php.net. Please make sure the mirror is fully functioning at all times and let us know about any planned maintenence so we can temporary disable the mirror. You are right that the manual alias check is the first one to run, so any outage is likely to be caught by that check. Currently both pl1.php.net and at2.php.net are enabled. -Hannes On Thu, Apr 30, 2015 at 5:41 AM, Oskar Szafraniec <[email protected]> wrote: > Hi Patrick, > > Same here. I received emai about a problem 2am. This is the time when backups are made and it can lead to short downtime of vm running pl1.php.net. I think manual-noalias check looks for a string in response. If there is no response or empty one, the check fails. If your vm and ours where offline at the time of check, we’ve received emails. > > Case solved ;) > > Best regards, > Oskar Szafraniec > > >> Wiadomość napisana przez Patrick Schaaf <[email protected]> w dniu 30 kwi 2015, o godz. 10:08: >> >> Hi Oskar, list, >> >> Am 30.04.2015 09:35 schrieb "Oskar Szafraniec" <[email protected]>: >>> >>> Dont’ know where is the problem. I’ve checked >> http://pl1.php.net/manual/noalias.txt and it displays „manual-noalias”. >> There is no „manual” alias in Apache config, manual.conf is not included at >> all. >> >> We had the same confusion yesterday with the at2.php.net mirror and a >> complaint about manual-noalias. At the time (due to hoster maintenance) the >> server was not reachable at all, and apparently the manual-noalias check >> was the first one to notice, test request failing, and misinterpreting >> that. Issue fixed itself once our mirror VM was reachable properly, again. >> >> There were changes to the checking infrastructure a few days ago, which >> probably cause this change in checking / reporting behavior. >> >> best regards >> Patrick >