Microsoft SQL Server High-Availability Videos and Q&A Log

You Want Videos? We Got Videos!

I always enjoy getting the chance to catch up with author, consultant, and Microsoft Clustering MVP Allan Hirt.  Allan and I recently presented two sessions covering an overview of high availability in Microsoft SQL Server and, the following week, a demo of how to implement several different kinds of high availability techniques including database mirroring, transactional replication, and Windows clustering services.

You can see videos of these presentations at the Quest Software Pain of the Week webcast site.

Archives of other videos are here.

We’re joined by my new cohort in crime, Iain Kick, also renowned for his role as the editor-in-chief at SQLServerPedia.

Allan is the author of these and other books:

During the Q&A, we always get a lot of questions.  Here are the collected Q’s and A’s.

~~~

Meeting Questions and Answers:
Subject: Pain of the Week We – October 14, 2010
Start Time: 14 October 2010 16:00:00 GMT+1:0
________________________________________
Paul T Asked: General comment….. copying files is where I love the SAN replication. Don’t trust it at all for real database stuff though!
________________________________________
Sherbaz M C P Changes Question To: Please suggest the best back/restore techique for huge databases. Should we go for SAN level backup technologies or snapshots that uses SQL Writer? Storage admins claims that they can restore a 28 TB database in 30 minutes. They are saying that they can even take transaction log backups and restores.
________________________________________
Ameena L Asked: When we talk to our management about SLA, Recovery object and uptime etc. They say just do your best. They expect no downtime but will not define any of this so how we determine what we are shooting for if target is not identified?
Kevin Kline Answered: That’s why actually working through a recovery is so important. I strongly recommend that you perform a full recovery, perhaps to a dev or QA server, and then using that benchmark as the general rules for your SLA.
________________________________________
Stuart A Asked: What do RTO & RPO stand for?
________________________________________
Fadel Asked: at 2 part questions: part1- what is the advantage of having a sql clustering over Oracle RAC where in the first case, we have at least 1 node sitting there doing nothing waiting on the other node(s) to fail, comparing to the second case where Oracle RAC offers both, HA and scalability – part2: what is the new features on sql 2008 clustering compared to sql 2005 clustering – thanks!
________________________________________
Mohammed Changes Question To: Shouldnt clustering considered as a local HA instead of DR since the other nodes are local?
________________________________________
Alan Xu Changes Question To: Is cluster 2008 support sql 2005 instance and sql 2008 instance in the same cluster?
________________________________________
sjohn Asked: What is the maximum recomended distance for a synced mirroring with a reasonable size of pipe
________________________________________
Anita Asked: You mentioned about multi – instance cluster( active-active)? I will be interested in learning more about this. we have active-active in our environment. Is it scheduled for next week?
________________________________________
robert l Asked: Will mirroring work with a large number of “Smaller” databases? 500 2gb databases vs 2 1tb database
________________________________________
Alvinator Asked: When log shipping if you have a large tlog will it ship it in chunks or the entire log file?:
________________________________________
James M Asked: What are the advantages of Database Mirroring over transaction replication? or vice versa
________________________________________
Martin Asked: IDEA – Maybe Quest should provide a Database Invenotry/Request to gather all the information for a HA environment
________________________________________
Martin Asked: I am very interested in finding a holistic view to HA and DR to handle SQL Server. Oracle and even DB2 – are you aware of such a product/s?
________________________________________
Alan Xu Changes Question To: How to handle SSIS in multiple instance?
________________________________________
Nadim Y Asked: I was coming from a consolidation angle – looking to install muliple instances on each node within the cluster. Is there an upper limit or just based on the resources available on the nodes
________________________________________
Mohammed Asked: IS clusteting possible in std edition sql 2008?
________________________________________
Anita Asked: You mentioned about multi – instance cluster( active-active)? I will be interested in learning more about this. we have active-active in our environment. Is it scheduled for next week?
Kevin Kline Answered: Next week is mostly demos of how to set up these techniques. However, there are some resource slides at the end that point you to more websites and whitepapers about each of these specific technologies.
________________________________________
robert l Asked: Will mirroring work with a large number of “Smaller” databases? 500 2gb databases vs 2 1tb database
Kevin Kline Answered: absolutely! however, there is some management overhead in that mirroring is configured on a per database level
________________________________________
Alvinator Asked: When log shipping if you have a large tlog will it ship it in chunks or the entire log file?:
Kevin Kline Answered: It doesn’t ship the transaction log, it shipps the backups/dumps of the t-logs. That’s an especially good use for backup compression. We have a lot of features specifically for this in our LiteSpeed backup tool.
________________________________________
James M Asked: What are the advantages of Database Mirroring over transaction replication? or vice versa
Kevin Kline Answered: We’re on those slides now. =^)
________________________________________
Martin Asked: IDEA – Maybe Quest should provide a Database Invenotry/Request to gather all the information for a HA environment
Kevin Kline Answered: I like it, Martin. I’ll put that in my notebook. Btw, have you seen our “Discovery Wizard for SQL Server”? It does inventorying for SQL Servers.
________________________________________
@SQLRich Asked: I’ve heard that not all of the SQL components are cluster aware, is this true?
Kevin Kline Answered: Yes, afaik, SSIS is one of the biggest offenders in this area. But I think there are others.
________________________________________
Desi Asked: Can you mirror a database to more than one locaton?
Kevin Kline Answered: I don’t recall the exact steps to do this, but yes. Also available with replication and log shipping.
________________________________________
Mohammed Asked: IS clusteting possible in std edition sql 2008?
Kevin Kline Answered: Discussed on slide 11, Mohammed. 🙂
________________________________________
sjohn Asked: What is the maximum recomended distance for a synced mirroring with a reasonable size of pipe
Kevin Kline Answered: That’s on slide 10, 🙂
________________________________________
Nadim Y Asked: Is there a maximum number of instances you can have in a multinode cluster
Kevin Kline Answered: Not sure what you mean, Nadim. Usually, I don’t see more than 2 actives to 1 passive.
________________________________________
Mohammed Asked: One question before we leave please..i asked this before also. Clustering is more Local HA and mirrroing/ls is DR…is this correct?
Kevin Kline Answered: Clustering across fibre can be miles, sometimes many miles distant, though it’s uncommon. Mirroring and log shipping can be long-distance. Latency varies per technology type and is also another issue as to whether the solution is considered HA or DR.

Speak Your Mind

*

This site uses Akismet to reduce spam. Learn how your comment data is processed.