httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stormy <>
Subject Re: [users@httpd] Unexpected shutdown
Date Sun, 03 Jul 2016 16:51:02 GMT
At 05:49 PM 7/2/2016 -0700, Kurtis Rader wrote:
>On Sat, Jul 2, 2016 at 5:26 PM, Stormy <> wrote:
>Apache/2.4.7 (Ubuntu 14.04LTS) Server built: Mar 10 2015 13:05:59; Intel 
>8-core 64-bit, 16 Gigs RAM.
>Perfect for 479 days.
>Crashed today (Saturday afternoon, no particular activity that I can 
>find.) Tried to restart # service apache2 reload, nothing... Error log 
>(nothing closely relevant before this):
>[Sat Jul 02 18:24:41.754237 2016] [mpm_prefork:notice] [pid 1673] AH00171: 
>Graceful restart requested, doing restart
>(98)Address already in use: AH00072: make_sock: could not bind to address 
>[Sat Jul 02 18:24:41.809755 2016] [mpm_prefork:alert] [pid 1673] no 
>listening sockets available, shutting down
>[Sat Jul 02 18:24:41.809758 2016] [:emerg] [pid 1673] AH00019: Unable to 
>open logs, exiting
>Rebooted (hate that) -- *seems* to be OK; hardware seems OK (LAN eth cards 
>listening, 31, 32, 33.) Nothing else in logs -- access shows 
>average use.
>We'll never know now.

Thank you (and Curtis, see below.) Most probably agree, but I'm still 

>If this happens again you'll need to employ tools like "sudo netstat -anp" 
>to find the process that has bound to port 80.

In fact, it was by running "# netstat -natp" that I found one eth card was 
down (and the rather problematic Ubuntu ifup|down didn't work, that's why I 
ended up "# shutdown -r".) The three other cards were OK, but the "normal" 
"tcp 0 0* LISTEN 1682/apache2" was missing.

I'm still trying to understand "[mpm_prefork:notice] ... AH00171: Graceful 
restart requested, doing restart" -- nobody was in the $workplace, only I 
have ssh, so who/what "requested"?  If it was a hardware failure (not 
obvious even for a single card), does Apache automagically try a graceful 
restart? -- even if three cards still showed "tcp 0 0|31|32:53* LISTEN 1412/named"? -- and what's the "(98)Address already in use"?

And At 09:00 PM 7/2/2016 -0400, Curtis Maurand wrote:
>It didn't fully go away. Some process was still there. Was it running fast 
>cgi or php-fm? Those processes moght still have been running.

No -- cgid was "enabled" (not fcgid), and nothing "php" (I'm a tad 
paranoid, and we only use Perl) ... [access_compat, alias, auth_basic, 
authn_core, authn_file, authz_core, authz_host, authz_user, autoindex, 
cgid, deflate, dir, env, filter, include, mime, mpm_prefork, negotiation, 
proxy, reqtimeout, rewrite, setenvif]

tnx and br -- Paul

View raw message