From: Haroon (itguy.pro@gmail.com)
Date: Wed Feb 18 2009 - 06:41:46 ARST
Great information and excellent points, Pavel, thank you. We are still in
process of planning this out so I will surely bring this post up again in
future.
Thank you much.
Haroon
On Sun, Feb 15, 2009 at 5:39 PM, Pavel Bykov <slidersv@gmail.com> wrote:
> Just a couple of pointers from the experience and from the top of my head:
>
> 1. You have to think in both directions and it will really depend on what
> you're failing over. If you have two sites with geographical server cluster,
> all you'll need is routing protocol convergence. Otherwise you might look
> into GSS. But with separate sites that are not balanced on application level
> where you would use DNS solution, it will not be "seamless" because of all
> the caches. What you could do is "anycast" by blocking/tracking
> availability, and in case of problems brining up same IPs in different
> location. But there are other implications to consider - like what if the
> problems are between two sites? You'll then have two IP addresses active
> basically balancing the traffic, which is not good unless servers are
> interconnected on application level, which brings us back to application
> cluster...
>
> 2. Uh, it really depends on the size, data criticality, server load etc.
> Also it depends if you are going to sync databases directly or Storage
> Systems. From experience 12 to 24 hours should be enough. But it really
> depends...
>
> 3. ACE (descendad of CSS) usually load balances very close to servers,
> therefore it's located close to server lan. It's purpuse is mostly to
> balance load locally, while GSS's purpose is to balance load geographically.
> It definitely could be used for failover as well, but for site failover to
> another site, the uses are more limited.
>
>
>
>
> On Thu, Feb 12, 2009 at 7:03 PM, Haroon <itguy.pro@gmail.com> wrote:
>
>> Hello Experts,
>>
>> My question is somewhat off topic, I hope that is okay. I am trying to
>> figure out how to setup site redundancy for online highschool and
>> university
>> setup, a backup site for our primary location so in case of
>> power/equipment
>> failure or ISP issues, the users are not effected by the down time.
>>
>>
>> 1. In case of failure at the main site, the switch over should be
>> seamless to the end user. I believe this is possible with DNS load
>> balancing, any recommendations? Off site dns load balancer?
>> 2. SQL databases should sync hourly or at least once a day for latest
>> data at the backup site.What software would do this?
>> 3. What type of equipment is required on each site to accomplish this
>> setup? Cisco's CSS of any use here?
>> 4. Do I need to get a dedicated site to site VPN?
>>
>>
>> Any other suggestion or ideas would be greatly appreciated.
>>
>> Thank you,
>>
>> Haroon
>>
>>
>> Blogs and organic groups at http://www.ccie.net
>>
>> _______________________________________________________________________
>> Subscription information may be found at:
>> http://www.groupstudy.com/list/CCIELab.html
>>
>>
>>
>>
>>
>>
>>
>>
>
>
> --
> Pavel Bykov
> ----------------
> Don't forget to help stopping the braindumps, use of which reduces value of
> your certifications. Sign the petition at http://www.stopbraindumps.com/
Blogs and organic groups at http://www.ccie.net
This archive was generated by hypermail 2.1.4 : Sun Mar 01 2009 - 09:44:11 ARST