NDO2DB Issue out of the blue

[BanditBBS]

It hasn't been an issue in 27 hours. Not since I did the make and make install on the debug version. Once I stopped that and put the original in and restarted its been running great every since. I'll of course update this on Monday again.

[BanditBBS]

45 hour mark and if failed...was just a good Saturday apparently. Nice thing though, I used the action component for the first time and create a script to fix(restart ndo and kill extra queues) it upon a click.

So, patiently waiting any feedback from the log, beta/alpha of a patch or any other idea.

EDIT: It has crashed 3 times this morning(Monday).

[jfrickson]

I looked over the log over the weekend and couldn't find anything suspicious.

Let's try something. Perform the same process as before, but instead of uncommenting all of the defines starting at line 72, just uncomment the first line from:

Code: Select all

/*#define DEBUG_NDO2DB 1*/                        /* don't daemonize */

to:

Code: Select all

#define DEBUG_NDO2DB 1                        /* don't daemonize */

Go to line 218. There you should see the following:

Code: Select all

	signal(SIGQUIT,ndo2db_parent_sighandler);
	signal(SIGTERM,ndo2db_parent_sighandler);
	signal(SIGINT,ndo2db_parent_sighandler);
	signal(SIGSEGV,ndo2db_parent_sighandler);
	signal(SIGFPE,ndo2db_parent_sighandler);
	signal(SIGCHLD,ndo2db_parent_sighandler);

at the end of that list, add the following lines:

Code: Select all

	signal(SIGHUP,ndo2db_parent_sighandler);
	signal(SIGILL,ndo2db_parent_sighandler);
	signal(SIGABRT,ndo2db_parent_sighandler);
	signal(SIGPIPE,ndo2db_parent_sighandler);
	signal(SIGALRM,ndo2db_parent_sighandler);
	signal(SIGUSR1,ndo2db_parent_sighandler);
	signal(SIGUSR2,ndo2db_parent_sighandler);
	signal(SIGBUS,ndo2db_parent_sighandler);
	signal(SIGPOLL,ndo2db_parent_sighandler);
	signal(SIGPROF,ndo2db_parent_sighandler);
	signal(SIGSYS,ndo2db_parent_sighandler);
	signal(SIGTRAP,ndo2db_parent_sighandler);
	signal(SIGVTALRM,ndo2db_parent_sighandler);
	signal(SIGXCPU,ndo2db_parent_sighandler);
	signal(SIGXFSZ,ndo2db_parent_sighandler);
	signal(SIGIO,ndo2db_parent_sighandler);
	signal(SIGPWR,ndo2db_parent_sighandler);

These are other signals that will cause the process to terminate.

Run ndo2db stand-alone as before, but you don't need to redirect stdout to a file. When the process dies, you should get a message saying:

Code: Select all

Caught the Signal 99 but don't care about this.

with the 99 replaced with some number. Let me know what (if anything) gets printed.

[BanditBBS]

Sure, thanks, I'll get right on this! (just need to figure out how to keep my terminal session active as it times out after inactivity - I'm sure there is a setting, lol).

[tmcdonald]

Sure, thanks, I'll get right on this! (just need to figure out how to keep my terminal session active as it times out after inactivity - I'm sure there is a setting, lol).

tmux or screen will at least keep the session active on the server so you can reconnect and not lose any work.

[rseiwert]

I'm wondering, if you set NDO2DEBUG directive and it does not daemonize, doesn't that stop it from using IPC and cause everything to run in one process?

[jfrickson]

I'm wondering, if you set NDO2DEBUG directive and it does not daemonize, doesn't that stop it from using IPC and cause everything to run in one process?

It does run in just one process, but it still uses IPC.

[BanditBBS]

Quick end of day update: Been running with modified ndo2db since 11:30ish(Central) and still running strong. Will update whenever it crashes.

[jfrickson]

These things never crash when you want them to. Thanks for keeping on top of it!

[BanditBBS]

These things never crash when you want them to. Thanks for keeping on top of it!

Nothing useful. It was at 100% and not doing its job so I can to close it manually, so its not crashing to give any code

EDIT: Went back to original binary. Its not crashing to give us anything, it just spikes to 100% and stops doing anything. Its crashed 2 more times since last night