Beefy Boxes and Bandwidth Generously Provided by pair Networks
more useful options

Re: Apache becomes unresponsive running mod_perl

by sundialsvc4 (Abbot)
on Jan 15, 2018 at 23:22 UTC ( #1207324=note: print w/replies, xml ) Need Help??

in reply to Apache becomes unresponsive running mod_perl

The very first thing that I do with a legacy application that uses mod_perl is to switch it to Plack.   The conversion process is actually very easy.

Plack uses the FastCGI mechanism to run the Perl code in a stable of separate worker-bee processes which can be managed by (Apache) itself.   The critical difference is that the Apache worker-processes are once again “lightweight,” and cannot themselves be ruined by a (Perl) application that doesn’t work, because the work is actually being done by a worker process that is held at arm’s length.   It is also very easy to offline-test the applications through the provided plackup command.

  • Comment on Re: Apache becomes unresponsive running mod_perl

Replies are listed 'Best First'.
Re^2: Apache becomes unresponsive running mod_perl
by stonecolddevin (Vicar) on Jan 16, 2018 at 00:27 UTC

    Plack uses way more than fastcgi, and a reverse http proxy is probably a lot easier than fastcgi

    Three thousand years of beautiful tradition, from Moses to Sandy Koufax, you're god damn right I'm living in the fucking past

Re^2: Apache becomes unresponsive running mod_perl
by Your Mother (Chancellor) on Jan 16, 2018 at 06:09 UTC

    Your answer, besides being almost completely wrong as usual (apache can be tested offline, PSGI is deployment agnostic so FCGI is an incorrect assertion, the processes are not lightweight, code conversion is not necessarily easy, and worker processes can be ruined or hanged, especially in an apache FCGI process, with application code), amounts to: "Change your technology stack completely." But at that point the error will not be solved. Deployment changes do not fix code issues. I'm really starting to wonder if the ones with valid lawsuits in all this, since it's a topic you like to revisit now and then, might really be the clients you've bilked with your advice.

      An accurate critique of the post in question, sadly nothing new. Don't expect a response from them however.

Log In?

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://1207324]
[marto]: good morning all, snow day here, won't make it to the office
[Corion]: marto: Oof - that means taking a day off or can you work from home?
[marto]: Corion, I can do some non-technical things, but there's no way for me to connect to the clients network.
[marto]: which is a shame, I had a really productive day yesterday
[Corion]: marto: Meh, so it'll be a day of cleaning out email...
[marto]: and hoped that I'd be able to continue the momentum :)
[marto]: I can't even access client email, nor my employers since those idiots moved to citrix
[marto]: it literally doesn't work. Also, via their citrix interface theres no way to open attachments, or upload an attachment for sending. The company policy is that you email you work account from a personal one, and forward it on from there :/

How do I use this? | Other CB clients
Other Users?
Others chilling in the Monastery: (5)
As of 2018-01-16 08:49 GMT
Find Nodes?
    Voting Booth?
    How did you see in the new year?

    Results (175 votes). Check out past polls.