What are the meanings of the various signals in PROCESS_SIGNALS?

Nicholas_Chamma

I'm trying to track down why one of our clusters went down. In the logs I see messages like this:

2014-05-21 03:24:02.322 Main:0x5b5f2b0 [Main] <INFO> Handling signal: 10

Looking at this page, I see no mention of what signal 10 means.

What are the meanings of the various signals in PROCESS_SIGNALS?

Evan_H

signal 10 is used to signal vertica to rotate logs.

Evan_H

 See if a new log is started for the vertica.log soon after signal 10. 

It could also mean bad memory access, but more likely rotate logs.

Nicholas_Chamma

OK, sounds good.

Where can I find a list of the meanings of all the possible signals?

What is signal 11, code 1, for example?

[Deleted User]

Hi Nicholas,

From your new thread elsewhere, it sounds like you may have figured this out -- these signals aren't Vertica signals; they are POSIX signals that are being received by Vertica.  You can find them documented, for example, here:

http://man7.org/linux/man-pages/man7/signal.7.html

(Scroll down to "Standard signals".)

Signals 10 and 12 are reserved for application-specific purposes; for example, things like log rotation.  Signal 11 is SIGSEGV, ie., invalid memory access.

Adam

Evan_H

http://man7.org/linux/man-pages/man7/signal.7.html

Was just about to reply the same.. :-)

[Deleted User]

Oops; sorry!  Was replying to the other thread; figured I should post here too.

Nicholas_Chamma

Yep, I eventually figured it out. I suggest updating the documentation on PROCESS_SIGNALS to indicate as much, because it's not obvious.

Nicholas_Chamma

Thanks guys!

Adrian_Oprea_1

I also think this should be documented !