The poor's man method: pipe the logs in a script, detect patterns of what
annoys you, send an snmp message (or a mail).
I used a similar technique to generate syslog events that were captured by TNG.
Below an excerpt from what I documented for my ops team (does not render well
in pure text - te
You can also take a look at the SRV record for DNS
(http://en.wikipedia.org/wiki/SRV_record) and pray for the browser to
support it.
From: Harsimranjit singh Kler [mailto:simran...@gmail.com]
Sent: Monday 6 February 2012 11:46
To: DENIS Laurent
Cc: users@httpd.apache.org
Subject: R
his first link:
http://www.cyberciti.biz/faq/rhel-centos-fedora-keepalived-lvs-cluster-c
onfiguration/
From: Harsimranjit singh Kler [mailto:simran...@gmail.com]
Sent: Monday 6 February 2012 10:53
To: DENIS Laurent
Cc: users@httpd.apache.org
Subject: Re: [users@httpd] High
If you just want failover on the head apache: use two nodes with vrrp +
keepalived (if you're on linux) to monitor your service.
If you need load-balancing, then you should use two VRRP addresses with
DNS round-robin. But you'll stumble on client-server persistence problem
if you don't use tomc
Enable the %D in the logs - it will tell you how long the server takes
to serve the query, in microseconds. It can be very long from client
side but fast from server side - and even if slow from server side, it
can be the client that hangs in the middle of the transaction (for very
large ressources
Hello,
*** Question ***
How can I measure how much time is spent in the processing function of
module?
Of course, the supplier of the module does not provide any source code.
My idea is to add an APRLOG_DEBUG the "apr_time_now" before and after
the function is called (don't know yet which hook
