I'm just in the process of taking over the ongoing support and maintenance of a series of Umbraco 4.7.0 sites. The development wasn't ours, and theres no documentation.
The client has been reporting recurring 503 Service Unavailable errors. I'm looking for help to diagnose these 503 errors.
The sites are hosted from a load balanced cluster of servers. Two web servers, one admin web server, and two database servers. The cluster hosts 5 sites each in their own umbraco instance & associated database. Four are relatively low trafficked and one which receives a much larger amount of traffic (70k visits per month).
The admin server is configured to send distributed calls to the two web servers. Files are shared from the admin server to the webservers using a file replication app called Mirror Folder, which pushes files from the following folders
/css
/masterpages
/media
/python
/scripts
/usercontrols
/xslt
across to the two web servers via windows shares they expose to their c drives.
I'm not seeing any confirmation of the 503 errors in the httperrXXX.log files. There's a lot of warning's in the event viewer of both web servers, typically in the form:
Application pool 'XXXXXXXXXX' is being automatically disabled due to a series of failures in the process(es) serving that application pool.
A process serving application pool 'XXXXXXXXXX' terminated unexpectedly. The process id was '4972'. The process exit code was '0xc00000fd'.
A process serving application pool 'XXXXXXXXXX' terminated unexpectedly. The process id was '4172'. The process exit code was '0xff'.
Looking in the W3C log files it looks like the app pool is restarting at least every half hour.
OK So it seems I was too hasty. The problem isn't solved. We had another 503 error this morning.
Since Saturday night, I've disabled Examine, and also tidied up some folder permissions as they weren't as I'd expect them. This has stemmed the tide of Examine errors reported in umbracoLog table. There are some errors in umbracoLog: At /umbraco/ping.aspx (Referred by: ): which looked to be caused by a missing ping.aspx file on the web nodes. So I've fixed that, but I don't really think it's got anything to do with it.
Does anyone have any ideas what could have caused this?
I am getting similar (but maybe not the same) with Umbraco 6 ... it appears to be caused when there is a problem in an XSLT file (even tho' it validates/saves correctly). This causes Umbraco to log an exception ("Error parsing XSLT file. Exception: System.OverflowException: Value was either too large or too small for an Int32.") which in turn seems to trip up log4net ("<log4net.Error>Exception during StringFormat: Input string was not in a correct format."). I suspect (tho' am not sure) that this causes some sort of cyclical error which crashes the app and disengages it from the app. pool (at least that's what happens in my case).
Recurring unexplained 503 errors
I'm just in the process of taking over the ongoing support and maintenance of a series of Umbraco 4.7.0 sites. The development wasn't ours, and theres no documentation.
The client has been reporting recurring 503 Service Unavailable errors. I'm looking for help to diagnose these 503 errors.
The sites are hosted from a load balanced cluster of servers. Two web servers, one admin web server, and two database servers. The cluster hosts 5 sites each in their own umbraco instance & associated database. Four are relatively low trafficked and one which receives a much larger amount of traffic (70k visits per month).
The admin server is configured to send distributed calls to the two web servers. Files are shared from the admin server to the webservers using a file replication app called Mirror Folder, which pushes files from the following folders
/css
/masterpages
/media
/python
/scripts
/usercontrols
I'm not seeing any confirmation of the 503 errors in the httperrXXX.log files. There's a lot of warning's in the event viewer of both web servers, typically in the form:
Bump
have you looked in umbracoLog table for clues ? (http://our.umbraco.org/forum/core/general/23589-Lucene-crashing-app-pool-on-DFRS)
#h5yr
OK So it seems I was too hasty. The problem isn't solved. We had another 503 error this morning.
Since Saturday night, I've disabled Examine, and also tidied up some folder permissions as they weren't as I'd expect them. This has stemmed the tide of Examine errors reported in umbracoLog table. There are some errors in umbracoLog: At /umbraco/ping.aspx (Referred by: ): which looked to be caused by a missing ping.aspx file on the web nodes. So I've fixed that, but I don't really think it's got anything to do with it.
Does anyone have any ideas what could have caused this?
I am getting similar (but maybe not the same) with Umbraco 6 ... it appears to be caused when there is a problem in an XSLT file (even tho' it validates/saves correctly). This causes Umbraco to log an exception ("Error parsing XSLT file. Exception: System.OverflowException: Value was either too large or too small for an Int32.") which in turn seems to trip up log4net ("<log4net.Error>Exception during StringFormat: Input string was not in a correct format."). I suspect (tho' am not sure) that this causes some sort of cyclical error which crashes the app and disengages it from the app. pool (at least that's what happens in my case).
is working on a reply...