September 10, 2006 at 3:39 pm
September 10, 2006 at 4:04 pm
I had SQL2K, SQL2005 and MySQL4.2 all coexisting on a development box.
I'm not sure if that is such a good idea on a production box but if you could have one box with 3 virtual servers on it that might be a better bet.
September 11, 2006 at 7:54 am
It really depends on the scope and usage of the servers. Go through the same questions you would ask yourself if you were setting up a second instance of SQL Server on that machine. I have 2k5 and MySQL 4.23 and 5.0 running on my PC at home for quite some time now and have had no issues with it. I would definately install on a dev machine first and do some load testing to make sure there are no issues. The other question would be, can you port the mySQL data to SQL Server and would that make you more comfortable with the arrangement, or does the data have to reside in mySQL?
EDIT: Spelling =)
September 11, 2006 at 9:45 am
Hi Todd,
There are several reasons why somebody might ask you to do this. Some are technical, some are not. Some are good (I don't mean to equate "good" with "technical" btw). Some are not good, but you might be wise to say yes to them instead of "that's not a good reason" anyway, depending on your situation.
Here are some I can think of right off the bat:
* -- The database is owned by a group/application/environment that works x-platform. If some of the client code is running on multiple platforms, they shouldn't have to put their data in SQL Server for their Windows instance while other instances are in MySQL. (This is an example of a really good reason. It doesn't matter to me whether the instances are test-production or production-production or some other scenario, FWIW.)
* -- There is a corporate policy to move to open source wherever possible, and this particular database is a new effort, so the idea is to start it off the way it will eventually go. (This is an example of "maybe not a good reason" but depending on your situation it may not be your place to argue with the policy.)
* -- The database is accessed by a lot of current client code that currently goes against a MySQL instance (in any environment). Management wants to "own" the data, which was originally set up by a user on an ad-hoc basis, but they need some level of comfort about how they consume the data and the process. There is some concern on the user's part, or on mgmt's, that a changeover to SQL Server in a hurry might break some client code. (If this is the reason, you might offer a compromise: create the migration process for them, add an incremental or refresh process for it so that the SQL Server version stays up to date, and give them time to adjust.)
If you need additional details, scenarios, please ask. If you simply need some coaching about dealing with MySQL because you haven't done it before, be honest with mgmt -- because it does represent a risk to be asked to consume a system and administer that is new to you, it's best to be clear about it. Be sure to be honest with yourself, too; not all your concerns are technical, but that's okay.
HTH,
>L<
September 11, 2006 at 9:50 am
There are a lot of reasons not to let anything (not just MySQL) be installed on your SQL servers. It will compete for memory, disk space, and processing power. As a DBA I'm a paranoid, greedy bastard as far as the servers are concerned, I don't want to hear about anything besides SQL on them. But most companies can't afford to buy a huge server for every application, and you might have one of the biggest servers available with lots of extra capacity. It might be a career-limiting move to not play nice and share.
You and I might not consider MySQL to be completely in the same league as SQL Server, but it does have some interesting features. You can choose a number of different database engines, and you may have a use for geographic indexes. Some workloads benefit from these features, so you have to admit the possibility that SQL Server is not the best solution in every case.
Is the request to load MySQL based on a desire to use some feature not supported by SQL Server? It would be hard to say no if it is something MySQL does better. Or is it just that some goofball read an article on MySQL and wants to play with it? Whether it is a good idea or not, if there is a lot of corporate political weight behind the request your best response may be to get on board and cooperate.
Do you have spare capacity on your server, or are you already hitting limits? You may want to document what tasks are loading your server already, and speculating on the impact of slowing SQL down. Will delays merely annoy some SQL users, or will crises occur as deadlines are missed? You should get some baseline performance measurements of your server before MySQL is installed. Get lots of permon and profiler traces, plus run times for critical jobs. Even if you are cooperating with the MySQL install you should be able to measure the impact of it. Any complaints you have later will be much stronger if you have numbers to back them up.
September 11, 2006 at 2:22 pm
Viewing 6 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply