I dropped a pair of pfSense routers (1.2-RC3) in front of a rather
large client base - revolving door of about 350 clients per hour and a
running throughput of ~8Mbps.  Everything went fine for a few hours,
then the lighttpd.error.log process started showing the following and
the whole system stopped passing traffic:

(mod_fastcgi.c.1731) connect failed: Connection refused on
unix:/tmp/php-fastcgi.socket-0
(mod_fastcgi.c.2885) backend died; we'll disable it for 5 seconds and
send the request to another backend instead: reconnects: 0 load: 194
(mod_fastcgi.c.3496) all handlers for  /indx.php on .php are down.
(mod_fastcgi.c.2703) fcgi-server re-enabled: uix:/tmp/php-fastcgi.socket-0

There were also numerous copies of this following as well, but were
happening before the failure and are occurring on the secondary as
well (with no seeming issue):

(network_freebsd_sendfile.c.97) writev failed: can't assign requested address 97
(connections.c.603) connection closed: write failed on fd 97

I failed over to the secondary, rebooted the primary (turned CARP off
on it), and everything seems to be running okay for now.  However,
looking in root's directory of the primary, there are several
'lighttpd-upload-*' files with roughly the same timestamps as the
issue occurring, look like iCal files.  What I presume happened was
that some Mac user connected and attempted to do a calendar sync
before authenticating; lighttpd/CP seems to have taken the uploads and
that precipitated some failure.  Any alternate ideas or ideas about
preventing this from happening again?  Of course, it really sucks that
the CP httpd is running as root.


RB

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to