SSO from moodle locks server 503 php_slow warnings

[hermannkm]

When doing many (15 pupils) simultaneous switches/logins from moodle to Mahara (Mahoodle installation, 3.0+ and 15.10.1) server locks until producing a 503 error and logging into php_slow.log:

[16-Dec-2015 08:42:29] [pool e48616] pid 13050
script_filename = /data/web/e48616/html/mahara/htdocs/auth/xmlrpc/land.php
[0x00007f1abf912228] curl_exec() /data/web/e48616/html/mahara-15.10.1/htdocs/lib/web.php:4224
[0x00007f1abf911f98] mahara_http_request() /data/web/e48616/html/mahara-15.10.1/htdocs/api/xmlrpc/client.php:55
[0x00007f1abf911b98] send() /data/web/e48616/html/mahara-15.10.1/htdocs/auth/xmlrpc/lib.php:103
[0x00007f1abf9119d0] request_user_authorise() /data/web/e48616/html/mahara-15.10.1/htdocs/auth/xmlrpc/land.php:78

These messages are repeated for many processes and make the two pages (Mahara/Moodle) unavailable for minutes. Most of the time a Apache 503 error ist produced:

83.218.163.159 - - [16/Dec/2015:08:44:06 +0000] mahara.klocker-mark.at "GET /auth/xmlrpc/land.php?token=8f469573ff0d075659fe0184f07348dd4722a5e7&idp=http://moodle.klocker-mark.at&wantsurl= HTTP/1.1" 503 0 "-" "Mozilla/5.0 (Windows NT 6.3; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/47.0.2526.80 Safari/537.36"

The system is practically unusable with multiple pupils at the same time.

I don't know exactly the server versions as I am on a remote server.

[Meg Goodine (meg-goodine)]

Is there any update on this issue? We are about to upgrade to Moodle 3.0 and Mahara 15.10. Would really appreciate knowing if others are also seeing this.

[hermannkm (h-klocker)]

Hi Meg,
there are good news: this problem seems to have gone away. It did without changing moodle or mahara - so I do not think it's their fault. The not so good news: it took me and my ISP to change some details of my account (they did not tell me which) to make this work.

[hermannkm (h-klocker)]

Just got notification about changing to invalid. This is OK with me ( was reporting this) as I think now this only happens at this specific provider with his setup. Changed to another provider -> could not reproduce this.