Fastcgi application not processing request (2nd attempt)

Apologies for sending this request twice - it was pointed out that my
original request was not very clear, so I have rewritten it to be more
meaningful (hopefully).

We are using mod_fastcgi-2.4.2 and Apache-2.0.52 with mod_SSL under RedHat
9.0 (2.4 kernel) to run a series of applications written in C using
fcgi_stio.h.

There is an intermittent problem, occurring about once in every 50 to 200
calls to one particular application where the HTTP headers are returned with
an empty body. The apache log shows the request succeeding with a 200 result
code.

The application normally writes to the log file as soon as it has accepted
the fastcgi request and before doing any processing. When the failure occurs
nothing is written to the application log or to the Apache error log;
processing does not occur.

Apache and thus fastcgi are logging at the info level.

The problem appears to be more likely to occur after the application has
been inactive for a long period e.g. at least a day, however, other fastcgi
managed applications successfully run immediately prior to the problem
application. The server and Apache are not restarted during the inactive
period.

We have observed and recorded this problem on three occasions with this
particular application, but it may not be limited to this particular
application. There is anecdotal evidence of other earlier failures with
similar results observed by the client application which may have been due
to one of the other server applications failing in the same manner.

According to the fastcgi docs, raising the logging level above info does not
give any extra information. How can we get more information out of fastcgi?

Any other ideas on how to diagnose and fix this problem will be appreciated.

Regards,

Rob.



___________________________________
fastcgi-developers mailing list
http://fastcgi.com/fastcgi-developers/