Dave Howorth wrote:
I agree with the things that you say but as I understand Per's situation, they don't fit his symptoms.
His problem on the production system apparently occurs with a single user making a single request - the first request. So the problem isn't load-related or complex-query-related.
Precisely. It is also reproducable on subsequent requests, although they seem to process slightly faster.
Per, various suggestions:
I do think you need a test system (the customer should pay really).
Yes and yes. Assuming I can come up with the proof needed, I will be sending the customer a bill for my time. Well, some of it :-(
I think asking a question on the processwire forum might well be useful.
Yep. I did in fact register for the forum yesterday, but have yet to receive the confirmation.
I think it would be useful to understand the difference between the developer's system and the production/test system. Can you or he run exactly the same strace test on his machine and get the same behaviour? If not, what is different?
I expect he's been working on Windows or Mac, but I think it ought to be enough to copy the website to some similar hardware and then compare numbers.
I'm not a PHP programmer and knew nothing about processwire (though I do know Perl). There seems to be a debugger for PHP called xdebug, and it seems to be possible to run processwire from the command-line, which makes using a debugger much easier. So if it was me, I would try to run the same request that you used for the strace from the command line with the debugger and single step it to find out where the lstat's are happening.
Interesting idea about running processwire on the command line, thanks. -- Per Jessen, Zürich (16.8°C) http://www.dns24.ch/ - your free DNS host, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org