RE: local search support
| From: | James Cox | Date: | Sat, 24 May 2003 14:00:02 +0000 |
| Subject: | RE: local search support | ||
| References: | 1 | Groups: | php.mirrors |
| Request: | Send a blank email to php-mirrors+get-18011@lists.php.net to get a copy of this message | ||
I missed this thread and discussion since I was out yesterday, however there
are a few things I would like to throw in here just to make things a bit
clearer.
> The roots of the problems are coming from the 'new way' or working here
> as it seems to me. I think this is visible since a few days here, and
> Wez's big changes are the crown on that. The following what is I see
> from the situation. I know Wez said he don't care what is visible 'from
> outside', but every member of the mirror maintanence group, who is not
> on systems@ sees things from outside.
>
this is unfortunate. However, there is no real way of fixing this; systems@
discussion must remain private for obvious reasons, and sometimes we run
into problems like we did just recently where we have to fix it, no matter
what. BTW, if this required having to rewrite a bunch of php engine code
because it was sloppy to fix it, we would, and it'd get checked in. That
would be commit-then-review, and there really is no problem with that
sometimes. In this instance, we didn't have time to email php-mirrors and
suggest the idea and discuss for a day or two. it wasn't feasible. we had to
either shut down www.php.net or just do it.
FWIW, the load has been at about 5-10 on a 5 minute avg for about 3 months,
since alot of the changes to the website were made. We have been
experimenting with various cache options to improve everything (we have been
running at 80%+ cpu for a while) and increase responsivity. We have been
experimenting with new kernels, reverse proxies, and god knows what else.
if this situation occurs again, about any service we provide, and we need to
make a change, we'll make that change. we can discuss later, and refine. I'm
sure Wez doesn't mind if we end up refactoring his code into something
else.... it did the job that was required.
As far as my recent posting of errors -- had I time to fix them, I would
have done... i did what anyone did -- posted them here and asked if anyone
had time to fix them. if no-one did, then that's fine... but that's the
spirit of community and co-operation. As far as Andrew's commit: well, that
was not an emergency commit, so the excuse of no-time for discussion doesn't
make sense :). There are also real reasons why it was the other way too, so
jumping in and committing was a bit wierd.
Gabor, don't take offence at all this. you aren't slaves. it's just that
sometimes we have to look in another direction, ie, looking after the
servers.
-- james