November 4, 2010 at 8:29 pm
Comments posted to this topic are about the item Enterprise Management
November 4, 2010 at 9:22 pm
Most tools I've come across charge per server for monitoring which is fine for a couple of servers but if you're running a large number becomes expensive quickly.
In the end I just developed my own system in .Net which loops through all the servers every X minutes and picks up any failing jobs, long running jobs, checks the SQL Error Logs for new issues, monitors disk space and alerts me to any servers that are down all presented in a nice GUI frontend.
It's not perfect but it works well for me and lets me know quickly when things go bad, it's also nice to have something you can build on whenever you have a few spare minutes.
November 5, 2010 at 12:02 am
Even with just 7 SQL Server instances to administer (as the accidental DBA), I felt the need to write my own tools.
I created a task in each server's maintenance plan to collect drive free space info into a local table after each backup. It also collects the size of the MDF & LDF files for each database.
My central DBA database has a list of all SQL servers to monitor and a VB.Net application I run every morning polls the other instances free space and log/data file sizes as well as backup info into a table that is displayed in a grid.
At a glance, I can see if backups ran (missing server = missing backup) and if a database log file grows unexpectedly, I know I need to check why. A separate display tells me if I need to worry about running out of free disk.
It may be basic, but saves a lot of time otherwise needed to log on to each server and check those issues.
November 5, 2010 at 3:01 am
What I liked was SQL Sentry (http://www.sqlsentry.com/) when I saw a demo four years ago. It's not that it just notices the problems but to allow to see where bottlenecks happens even time after when it happened. So it allows to check why all of a sudden last night the backup stalled or which query caused the bottleneck.
Back in the days when I had to manage my server I had a simple job log and each job just logged itself with start and end time. The whole thing put into a nice report (Business Objects DeskI at the time) nicely showed when bottlenecks started during the BI data load (from North America and Europe). I remember once I had to move a job 5 minutes earlier so that the running time dropped from 2h to 1h. Without this report it would have been hard to spot the creep in the later job causing the problem.
November 5, 2010 at 5:37 am
About 100 instances and a big user of MOM (soon migrating to SCOM). You can add your own scripts to it so we do extra things like look for things that have NOT happened, rather than just failed which MOM out of the box concentrates on. This means databases not backed up, when agent jobs last run.
We also use Athene for capacity and performance stats.
---------------------------------------------------------------------
November 5, 2010 at 6:55 am
Our failed db backups email us. We use BMC Sitescope product that monitors MSSQLSERVER and SQLServerAgent services and email us when they are down. I manage about 40 prod SQL Servers with about 300 diff databases. It is starting to get out of hand and more new SQL Servers and databases going to production every month.
November 5, 2010 at 7:15 am
We utilize SSIS and the "Server Overview" articles written here (now I have to go look it up so I can post the link). A job fires off this package, which hits all servers, collects data, pushes reports to an Excel spreadsheet, and email's us the link to the file which is waiting for us in our inbox when we arrive in the morning. With about 35 servers, it takes about 15 minutes to review all of the servers.
Wayne
Microsoft Certified Master: SQL Server 2008
Author - SQL Server T-SQL Recipes
November 5, 2010 at 7:24 am
I've used MOM and SCOM, as well as CMS with Policy Based management for some reporting aspect.
Unicenter, Netbackup, I3, and a little PDW.
I like Policy Based Management and SSIS, and I'm looking at doing some work with the EPMF soon.
http://epmframework.codeplex.com/
Twitter: @SQLBalls
Blog: http://www.SQLBalls.com
Channel: https://www.youtube.com/@Tales-from-the-Field
November 5, 2010 at 7:30 am
Since our ERP is Oracle and we had purchased the unreasonably expensive Grid Control (nee "OEM") license packs, I evaluated the SQL Server management plugin for Grid Control. It's OK for alerts, but I wouldn't want to manage a DB/server that way. And for $1800/CPU list fee (plus $400/yr maint!), that's waaaay outta line IMNSHO. So I wrote my own SQL Server plugin for Grid Control. Right now, it's very basic, giving me a config report and up/down alerts, but the price is right. 🙂
For day-to-day management on our ~30 small SS DBs, I use Quest's Toad for SQL Server. The Log Reader in Toad is fantastic and worth the price for that alone for this solo DBA. It also has a builtin feature where I can run SQLs against any of the DB connections I've made, which makes it a snap to run my management reports and have them saved in Excel.
Rich
November 5, 2010 at 7:50 am
Unicenter i$ good to Monitor Harware and $erver health.
$potlight on SQL Server by QUEST is good, but I liked the version before they made it look more like a stoplight than a spotlight.
SQL Sentry Performance Advisor for SQL Server looks like a great tool, but I have not used it personaly.
If you want to Role your own solution, Share Point and SSRS are good.
To be honest though, if I had to pay for something out of my pocket, I would get the SQL Monitor by red gate.
However, in my environment I need Hardware and System level monitoring tools that support email and text message alerts. Along with availablity reporting and a World Map screen for the VIP's and Board Room. Our solution also needed to include support for the network equipment, SAN specific hardware, server room environmental controls, and telco hardware.
These liability monitoring requirements limit the products available that can be used.
November 5, 2010 at 8:19 am
Two of us manage 160 instances, and 660 user databases with home grown tools. SQL Agent jobs on the instances do realtime reporting to a central instance - backups, backup audits, index work, daily collection of storage, logins, userids, instance stats, database stats, all o/s Event Log errors, anything over threshold, and so on. Emails are sent on hot events such as alerts. And we keep SSMS Query Windows connected to that central instance, and regularly query the realtime data.
We have done nothing exceptional, except maybe for the discipline of good documentation that we follow to the letter, which has resulted in servers and instances which are all very similar. We spend several hours getting a new instance set up correctly, but then never need to log on to that instance or server, unless an issue has been flagged.
November 5, 2010 at 9:40 am
SQL Response by Redgate and SQL DM by Idera are only good if you have a few servers. otherwise they are just too expensive to justify to managment. I use SQL Overview which Wayne mentioned already, it's free, runs from one central place, and gives me most of the critical info I need to know about on a daily basis for a lot of db servers. Here is the link for the entire package: http://www.sqlservercentral.com/articles/Integration+Services+(SSIS)/61774/ 😀
"Technology is a weird thing. It brings you great gifts with one hand, and it stabs you in the back with the other. ...:-D"
November 18, 2010 at 5:49 am
It's interesting that people monitoring a relatively few number of servers use commercial software, but those with large number of servers (in the hundreds) can't because of the cost! I realize software companies need to make money, but what about considering a type of Enterprise license that would allow companies that need a large number of licenses to be able to monitor their large environment without an unreasonable financial outlay. The monitoring software companies would still be pulling in revenue that they are now completely losing. Companies are looking for cost effective ways to 'buy not make' but cannot justify a monitoring product that charges by the processor!
November 22, 2010 at 10:33 am
I have used SSIS packages to import information from all of the instances and then agent jobs to run the package as well as fire off particular alerts.
I need to get my hands into Central Management and some other tools.
Jason...AKA CirqueDeSQLeil
_______________________________________________
I have given a name to my pain...MCM SQL Server, MVP
SQL RNNR
Posting Performance Based Questions - Gail Shaw[/url]
Learn Extended Events
Viewing 14 posts - 1 through 13 (of 13 total)
You must be logged in to reply to this topic. Login to reply