Simon icon Simon
Flexible server monitoring

Bogus HTTP/HTTPS test failures, Part 2

The first thread about this bug got locked/closed, so I'm starting a new one. :-)

So I'm on the 3.2 beta (3.2b4, 3012b), and that was the thing that was hopefully going to help to identify the source of this super-intermittent bug. Just got one today, and below is what the Console showed, hopefully enough to finally nail this one down....

Thanks,
John

8/17/11 2:51:05 PM Simon[225] NSUncaughtSystemExceptionException: Uncaught system exception: signal 10
8/17/11 2:51:05 PM [0x0-0x1a01a].com.dejal.simon[225] head: illegal line count -- +-1
8/17/11 2:51:05 PM Simon[225] An uncaught exception was raised
8/17/11 2:51:05 PM Simon[225] Uncaught system exception: signal 10
8/17/11 2:51:05 PM Simon[225] NSUncaughtSystemExceptionException: Uncaught system exception: signal 10
8/17/11 2:51:06 PM [0x0-0x1a01a].com.dejal.simon[225] 1 CFRunLoopRunSpecific (in CoreFoundation) + 3792
8/17/11 2:51:06 PM [0x0-0x1a01a].com.dejal.simon[225] 2 +[NSURLConnection(NSURLConnectionReallyInternal) _resourceLoadLoop:] (in Foundation) + 284
8/17/11 2:52:21 PM Simon[225] Pantomime: Connected!
8/17/11 2:52:22 PM Simon[225] Pantomime: Sending the message...
8/17/11 2:52:22 PM Simon[225] Pantomime: Sent!
Closing the connection.
8/17/11 2:52:31 PM Simon[225] Pantomime: Connected!
8/17/11 2:52:32 PM Simon[225] Pantomime: Sending the message...
8/17/11 2:52:32 PM Simon[225] Pantomime: Sent!
Closing the connection.
8/17/11 2:54:22 PM Simon[225] Pantomime: Connected!
8/17/11 2:54:23 PM Simon[225] Pantomime: Sending the message...
8/17/11 2:54:23 PM Simon[225] Pantomime: Sent!
Closing the connection.
8/17/11 2:54:32 PM Simon[225] Pantomime: Connected!
8/17/11 2:54:33 PM Simon[225] Pantomime: Sending the message...
8/17/11 2:54:33 PM Simon[225] Pantomime: Sent!
Closing the connection.
8/17/11 2:56:27 PM Simon[225] Pantomime: Connected!
8/17/11 2:56:27 PM Simon[225] Pantomime: Connected!
8/17/11 2:56:29 PM Simon[225] Pantomime: Sending the message...
8/17/11 2:56:29 PM Simon[225] Pantomime: Sending the message...
8/17/11 2:56:30 PM Simon[225] Pantomime: Sent!
Closing the connection.
8/17/11 2:56:30 PM Simon[225] Pantomime: Sent!
Closing the connection.
8/17/11 2:56:30 PM Simon[225] Pantomime: Connection closed.
8/17/11 3:00:23 PM Simon[225] Pantomime: Connected!
8/17/11 3:00:24 PM Simon[225] Pantomime: Sending the message...
8/17/11 3:00:25 PM Simon[225] Pantomime: Sent!
Closing the connection.
8/17/11 3:00:33 PM Simon[225] Pantomime: Connected!
8/17/11 3:00:34 PM Simon[225] Pantomime: Sending the message...
8/17/11 3:00:35 PM Simon[225] Pantomime: Sent!
Closing the connection.
8/17/11 3:01:28 PM Simon[225] Pantomime: Connected!
8/17/11 3:01:28 PM Simon[225] Pantomime: Connected!
8/17/11 3:01:30 PM Simon[225] Pantomime: Sending the message...
8/17/11 3:01:31 PM Simon[225] Pantomime: Sent!
Closing the connection.
8/17/11 3:01:31 PM Simon[225] Pantomime: Sending the message...
8/17/11 3:01:32 PM Simon[225] Pantomime: Sent!
Closing the connection.

David Sinclair's picture

Re: Bogus HTTP/HTTPS test failures, Part 2

Yeah, my site automatically locks comments on older topics after a while, to minimize the targets of spammers.

I'm sorry this problem recurred; I had really hoped the changes in 3.2 would fix it.

The exception seems to be way down deep in the OS HTTP handlers, so I'm not sure what else I can do to fend it off, other than moving the Web plugin into a helper app, which is something I do plan to do at some point. I was considering doing that for 3.2, but deferred it to 3.3 in the hopes the other changes would be enough.

But I'll take another look at the code and see if there's anything else I can do in the meantime.

JohnDCCIU's picture

Re: Bogus HTTP/HTTPS test failures, Part 2

Is it possible that this is a PPC bug and that Intel machines never experience it? I'm wondering if any other reports that you've gotten of this behavior are from Intel boxes: if they're all PPC and it's a PPC only bug, I can accelerate my move to Intel for Simon just to get this bug out of my hair....and with PPC being as old as it is, maybe that's the solution for the others that are experiencing this, if it's truly PPC-related.

David Sinclair's picture

Re: Bogus HTTP/HTTPS test failures, Part 2

I'm not sure which processors others have, but that's an entirely plausible theory. I know PPC is a lot more sensitive to memory issues than Intel, e.g. it will have garbage values in uninitialized variables, but Intel tends to have tidy zero values.

As mentioned in a recent blog post, I do plan to drop PPC support for Simon around the end of the year, so that's another justification in upgrading.