On 17/03/2015 07:59, Alfredo De Luca wrote:
Hi all.
Any clue on this?
On 15/03/2015 9:30 PM, "Alfredo De Luca" <alfredo.del...@gmail.com
<mailto:alfredo.del...@gmail.com>> wrote:
Hi all.
I ve never done this before so I am asking best practice/info/docs of
how to have 2 apache web servers in load balancing.
- Which httpd module do I have to load in the http conf?
Nothing unless you are using apache load balancer modules as a front end
... (mod_proxy_balancer)
- I was reading that I have to have a web load balancer on top of
them? Is it necessary? Can they accept requests from a cisco /F5 load
balancer?
I haven't played with F5 load balancers - but use the rival product
Brocade/SteelApp/StingRay/Zeus
traffic managers - which I think the F5s do the same thing as (just not
as user friendly)
so they should be able to do the job (I know we looked into it when we
bought the ZTMs)
- What about persistent connection?
? HTTP is stateless - if you have poorly written backends which require
requests to go the backend
you should be able to use sticky sessions - but this is bad as you lose
resilience (one of the main
reason for load balancing backends!)
- Also we''ll have a mySQL server? Any more info about this?
Load balancing MySQL can be trickier - easier if mainly RO connections
(you can round robin requests
to a large number of clones - or usually slaves to a single master) but
harder if read/write
- you can look at mysql cluster or master-master MySQL (galera)
Read write you can use master + multiple slaves - but need to tag a
process/session/user as
requiring access to master if a write happens for an unspecified length
of time!
Thanks in advance
--
Alfredo
---
This email has been checked for viruses by Avast antivirus software.
http://www.avast.com
--
The Wellcome Trust Sanger Institute is operated by Genome Research
Limited, a charity registered in England with number 1021457 and a
company registered in England with number 2742969, whose registered
office is 215 Euston Road, London, NW1 2BE.
