Re: inn 2.7.0, expire takes ages and hangs

Matija Nalis <mnalis-news@voyager.hr> wrote:

It just waits on select, and does nothing else.

 
Usually I'd use lsof(8) to see what is on the other side of that socket, and
try to continue debugging there.

Good idea, I forgot about this possibility, and I wanted to bring server
back up quickly.

1. Is it possible that I messed anything up by forcefully killing it?

 
Was it regular TERM or KILL (-9) signal?

TERM, expire process accepted it.

2. Why did it hang / how can I diagnose it / can I diagnose it?

 
you can add "-v level" to increase verbosity of expire(8).

Nice. I just read the manual (should've done it earlier...), now it's more
clear.

Another question: why does it expire anything if I have expiration
disabled?

 
And how exactly did you have "expiration disabled"?

I mean the article expiration, not history. I use CNFS, so articles expire
only when buffers roll over, and I don't forcefully expire my timehash (at
least for now). So the only expiration I'd expect is cleaning the history
(but probably these statistics are also about it).

For some reason I was convinced that expire only expires articles, and
expireover takes care of the overview AND history. Which doesn't make
sense, as overview can be disabled on transit servers, but they still need
history...

Or do you have reason to believe otherwise?

No, it's clear now.

So, it seems I still need expire to trim the history.

I think I'll make the copy of the server (I'll just copy the running
instance, it might be inconsistent but it won't matter), reconfigure its
feeds (so my server feeds the copy), and experiment on a copy. Lots of
data to transfer over the Internet, but I don't have a better idea on how
else to diagnose it...