Symptoms

Sometimes OA Control Panel does not load a page and shows the message:

Service Unavailable. The corresponding service is not available at the moment, try it within 2-3 minutes.

ssl_error_log on the branding host contains the errors during the outage:

[Mon Jul 11 07:29:42 2016] [error] [client 199.115.107.10] [Mon Jul 11 07:30:12 2016] [error] [client 199.115.107.10] (70007)The timeout specified has expired: proxy: error reading status line from remote server 192.168.50.208, referer: https://controlpanel.provider.com/single.html
[Mon Jul 11 07:29:42 2016] [error] [client 199.115.107.10] [Mon Jul 11 07:30:12 2016] [error] [client 199.115.107.10] proxy: Error reading from remote server returned by /servlet/Turbine/frm/single/, referer: https://controlpanel.provider.com/single.html

At the same time, UI part starts building the page, and it takes a lot of time eventually:

2016-07-11 07:29:42,891            c-123 DEBUG  OUT                  PAGE BUILD STARTED
2016-07-11 07:30:31,551            c-123 DEBUG  OUT                  PAGE BUILD FINISHED: [48660 ms]

Within this session, there are methods that take longer than expected, for example:

2016-07-11 07:29:42,894            c-123 DEBUG  CORBA                #583476: Plesk._hubStub.acquire_own_cred(undefined) >>> entry
2016-07-11 07:30:13,240            c-123 DEBUG  CORBA                #583476: Plesk._hubStub.acquire_own_cred(undefined) <<< exit [30346]

2016-07-11 07:30:13,252            c-123 DEBUG  CORBA                #583488: Plesk._BrandStub.getBrand(user_id=6 (su_user_id=6)) >>> entry
2016-07-11 07:30:31,442            c-123 DEBUG  CORBA                #583488: Plesk._BrandStub.getBrand(user_id=6 (su_user_id=6)) <<< exit [18190]

The calls are processed slowly on the OA MN backend.

Cause

Database server slowness or deadlocks cause simple methods get processed longer than usually.

Resolution

In case the issue happens frequently, investigate the reason behind the slow queries and eliminate it.

If the issue is rare, it is enough to raise Apache HTTP timeout on all hosts, serving CP requests:

[root@ng01 ~]# grep ^Timeout /etc/httpd/conf/httpd.conf
Timeout 60

# service httpd restart

and avoid occasional failures during page loads.

Internal content