Hi,<br>Amazon EC2 instances plus their EBS disks.<br>Scalable, available, reliable (from my experience) and you can experiment for a few $'s a day.<br><br><br><div class="gmail_quote">On Wed, Mar 4, 2009 at 11:06 AM, Ira Abramov <span dir="ltr"><<a href="mailto:Lists-Linux-IL@ira.abramov.org">Lists-Linux-IL@ira.abramov.org</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">related to the Citrix-vs-VMware question, in the spirit of the times...<br>
<br>
I want to create a way to host a cheap HA solution for a web hosting<br>
outfit. they are running a few pretty busy asymetric servers and want to<br>
start improving that infrastructure. each machine holds several dozens<br>
of virtual hosts.<br>
<br>
At the moment each server has its own local storage and mysql. every<br>
part is a SPOF other than the minimal RAID and such things.<br>
<br>
I'm thinking:<br>
* Move to a central non-virtual MySQL for the backend.<br>
* have two servers go P2V and have those VMs hosted back on their<br>
original hardwares (sadly this means some painful downtime), and find a<br>
way to let them crash-migrate for HA (still trying to figure this out)<br>
* Second stage, add a second MySQL in a master-master setup.<br>
<br>
I'd love to have two servers with symetrical setup, but as you can<br>
guess, the virtual hots are dozens of different apps that are too<br>
expensive to go and rewrite for clusters at this point, with the issues<br>
of user-uploaded files having to be available to both Apaches, etc.<br>
<br>
Assuming we want the cheapest reliable solution, i.e. not a $6K-20K SAN<br>
and FC, I am looking for an easier solution (easier on the pocket at<br>
least). However NFS proved to be a disaster in such cases (high-load web<br>
services), OCFS has not been nice to me with any setup other than maybe<br>
Oracle clusters, and GFS also never ran smoothly in my tests.<br>
<br>
Also OCFS and GFS require a common disk, which at this budget would be a<br>
Linux machine running an iSCSI target at best (or OpenNAS).<br>
<br>
Am I missing something? Can this kind of reliability be achieved without<br>
shelling out big bucks?<br>
<br>
Of course, The other option is just separate the MySQL, have a third<br>
machine rsync the files of the two (non virtual) servers every few<br>
minutes and have the hosting farm's layer4 switch redirect to the<br>
fallback if something happens. Not very "smart" nor scalable, but does<br>
70% of what we need till a bigger investment is required.<br>
<br>
your thoughts, as before, are welcome...<br>
<br>
Thanks,<br>
Ira.<br>
<br>
--<br>
Can't catch me yet<br>
Ira Abramov<br>
<a href="http://ira.abramov.org/email/" target="_blank">http://ira.abramov.org/email/</a><br>
<br>
_______________________________________________<br>
Linux-il mailing list<br>
<a href="mailto:Linux-il@cs.huji.ac.il">Linux-il@cs.huji.ac.il</a><br>
<a href="http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il" target="_blank">http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il</a><br>
</blockquote></div><br><br clear="all"><br>-- <br>Constant change is here to stay!<br><br><br><br>