From: F. Behrens <webware-devel@sp...> - 2014-04-30 17:55:18
for the first time in many years, a webware project for a customer has
developed constant failure conditions. I do, however, have problems to
isolate possible causes.
Webware runs with mod_webware in apache. In irggualr intervals, the
Exception in worker thread
-- CMDERR: cmd: ['CONNECT', 669L, 0] -> reply: Connection already active.
Traceback (most recent call last):
File "/opt/webware/WebKit/ThreadedAppServer.py", line 688, in threadloop
---------> Duplicate connect requests from 126.96.36.199 and
File "/opt/webware/WebKit/ThreadedAppServer.py", line 1134, in
requestDict = self.receiveDict()
File "/opt/webware/WebKit/ThreadedAppServer.py", line 916, in receiveDict
block = self._sock.recv(missing)
error: [Errno 9] Bad file descriptor
[ ... 200-300 more of these tracebacks, then a lot of consecutive
"Exception in worker thread" lines, then more tracebacks and so on ...]
As I see it, the failure point is in AppServer, while trying to receive
parts of a request header in dictionary format - apparently from a
socket that has somehow become invalid. Interesting enough it can happen
that some hours(!) later, the AppServer suddenly continues to work
properly, without restart.
When I restart the Appserver completely, the problem is gone immediately.
I'd appreciate any hint on where to begin the investigation and what to
Thank you for reading!
From: Christoph Zwerschke <cito@on...> - 2014-05-01 12:43:37
Am 30.04.2014 18:32, schrieb F. Behrens:
> for the first time in many years, a webware project for a customer has
> developed constant failure conditions. I do, however, have problems to
> isolate possible causes.
Never came across this kind of error. Maybe there is a memory leak and
it happens your server is running out of memory? I would monitor memory
usage and look if this is correlated.