[sclug] Annoying thttpd problem

[J.Mann]

> My home server run thttpd, because it's small, lightweight, and doesn't take 
> 15 seconds to start up on my elderly hardware. I really like it.

I also really like it, for the same reasons.

> There is, however, one problem, which is that every so often it shuts down for 
> no apparent reason. It doesn't crash, but it doesn't clean up after itself 
> which means I have to manually remove the file in /var/run before I can 
> restart it.

I have never experienced this. How often does it occur?

> Most of the message is its hourly stats dump (which I'd like to turn off 
> because it's annoying).

Unfortunately this is not configurable. The Debian package is compiled to
dump the stats every hour. 

> Jan 19 06:34:17 tiar thttpd[8463]: bind 0.0.0.0 - Address already in use
> Jan 19 06:34:17 tiar thttpd[8463]: Could not perform initialization. Exiting.
> Jan 19 06:34:18 tiar thttpd[3028]: exiting

There are two different processes here. Perhaps a race condition, caused
by a logrotate?

> This is a Debian unstable system. (It did the same thing with stable.)

I assume you understand that by running the "unstable" version, you 
are not receiving security updates.

Regards,
Jon Mann.


On Wed, Jan 19, 2005 at 02:38:55PM +0000, David Given wrote:
> My home server run thttpd, because it's small, lightweight, and doesn't take 
> 15 seconds to start up on my elderly hardware. I really like it.
> 
> There is, however, one problem, which is that every so often it shuts down for 
> no apparent reason. It doesn't crash, but it doesn't clean up after itself 
> which means I have to manually remove the file in /var/run before I can 
> restart it. It leaves these log messages:
> 
> Jan 19 06:34:16 tiar thttpd[3028]: up 86541 seconds, stats for 141 seconds:
> Jan 19 06:34:17 tiar thttpd[3028]:   thttpd - 0 connections (0/sec), 0 max 
> simultaneous, 0 bytes (0/sec), 9 httpd_conns allocated
> Jan 19 06:34:17 tiar thttpd[3028]:   libhttpd - 125 strings allocated, 51222 
> bytes (409.776 bytes/str)
> Jan 19 06:34:17 tiar thttpd[3028]:   map cache - 40 allocated, 0 active (0 
> bytes), 40 free; hash size: 1024; expire age: 1800
> Jan 19 06:34:17 tiar thttpd[3028]:   fdwatch - 68 selects (0.48227/sec)
> Jan 19 06:34:17 tiar thttpd[3028]:   timers - 4 allocated, 3 active, 1 free
> Jan 19 06:34:17 tiar thttpd[8463]: socket :: - Address family not supported by 
> protocol
> Jan 19 06:34:17 tiar thttpd[8463]: bind 0.0.0.0 - Address already in use
> Jan 19 06:34:17 tiar thttpd[8463]: Could not perform initialization. Exiting.
> Jan 19 06:34:18 tiar thttpd[3028]: exiting
> 
> Most of the message is its hourly stats dump (which I'd like to turn off 
> because it's annoying). The last part is the unusual part.
> 
> I have no idea what's going on here. thttpd is being started from /etc/init.d, 
> I don't have an entry in inetd.conf, I don't see why it's trying to rebind 
> the address seeing as it's already running ---
> 
> Any insight to offer?
> 
> This is a Debian unstable system. (It did the same thing with stable.)
> 
> -- 
> +- David Given --McQ-+ 
> |  dg at cowlark.com    | "The further you are from your server, the more
> | (dg at tao-group.com) | likely it is to crash." --- Woodhead's law
> +- www.cowlark.com --+ 
> _______________________________________________
> sclug mailing list
> sclug at sclug.org.uk
> http://www.sclug.org.uk/mailman/listinfo/sclug
>