[LRUG] Feet up/hands off/phone off managed hosting/peace of mind

[Riccardo Tacconi]

I've been supporting a web site for three years with HA configuration on
Linode (with another guy in the Czech Republic helping me for the 24/7/365
support), the web site was not ready to run on Heroku at that time. I am
using Passenger with Mysql Galera replication on three nodes. I need at
least one node running to have the site up. If you need to restart Apache
it means that something is wrong and restarting it will not fix the issues.
Sometimes the problem could be not having enough RAM, or Passenger need a
bit of tuning. Having Pingdom is not a solution for monitoring, you should
monitor other parameters, CPU, RAM swap, disk. In the meantime if you need
something to be restarted if it goes down, you could use bluepill, but I
prefer fixing the issues :-)

This book is good for learning how to troubleshoot issues:
http://www.amazon.co.uk/DevOps-Troubleshooting-Linux-Server-Practices/dp/0321832043
.

Heroku, Brightbox or Engine Yard are good options.

On 24 September 2014 12:25, Frederick Cheung <frederick.cheung at gmail.com>
wrote:

> Also consider automation - automated failovers can be tricky to get right
> but will happen faster than any human based option. With a platform like
> heroku that's just part of the service, but it's easy enough to do for
> application servers on AWS with autoscaling.
>
> Database failovers can be trickier. Mongo pretty much takes care of itself
> if you use a replica set, for relation databases we use amazon's RDS which
> takes care of this (for the record we had a database failure around 4am on
> boxing day last year. Other than a flurry of airbrake errors for a few
> minutes I barely noticed)
>
> Fred
>
>
> On 24 September 2014 at 12:18:54, Tim Harding (tim.harding at gmail.com)
> wrote:
> > Yes! Though I think that managing a team that costs more than our revenue
> > would probably cause more stress than having to leave my phone on in case
> > Pingdom alerts. ;)
> >
> >
> >
> > On 24 September 2014 12:04, Don Werve wrote:
> >
> > >
> > > On Sep 24, 2014, at 19:58 , Gabe da Silveira wrote:
> > >
> > > > Another strategy to consider: hire at least one developer in Sydney,
> one
> > > in London, and one in San Francisco ;)
> > >
> > > This is one of the benefits of building a distributed team — on-call
> > > schedules just become part of the workday.
> > > _______________________________________________
> > > Chat mailing list
> > > Chat at lists.lrug.org
> > > Archives: http://lists.lrug.org/pipermail/chat-lrug.org
> > > Manage your subscription:
> http://lists.lrug.org/options.cgi/chat-lrug.org
> > > List info: http://lists.lrug.org/listinfo.cgi/chat-lrug.org
> > >
> > >
> > _______________________________________________
> > Chat mailing list
> > Chat at lists.lrug.org
> > Archives: http://lists.lrug.org/pipermail/chat-lrug.org
> > Manage your subscription:
> http://lists.lrug.org/options.cgi/chat-lrug.org
> > List info: http://lists.lrug.org/listinfo.cgi/chat-lrug.org
> >
>
> _______________________________________________
> Chat mailing list
> Chat at lists.lrug.org
> Archives: http://lists.lrug.org/pipermail/chat-lrug.org
> Manage your subscription: http://lists.lrug.org/options.cgi/chat-lrug.org
> List info: http://lists.lrug.org/listinfo.cgi/chat-lrug.org
>



-- 
Riccardo Tacconi

http://github.com/rtacconi
http://twitter.com/rtacconi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lrug.org/pipermail/chat-lrug.org/attachments/20140924/143d3e81/attachment.html>