January 14, 2015 at 8:07 pm
I Asked this question from another forum :
I got a dba vacation in new office.
Please say me first and important steps for dba in new Office.
Whitch imfomation , I should get and what are most periority tasks.
thank you.
And they answered me : (This is ok)
1- I start by looking at the jobs, checking how maintenance type tasks are handled.
2- I checked the Error Logs to see if there are any immediate issues.
3- I check the load on the server. I get a sense for how busy it is, checking blocking, wait stats, PerfMon, etc.
4- I check database properties, sp_configure, LPIM and other settings. I check how alerts are configured.
But I need more detail and practical steps, about step 3 and 4 .
Thankyou
Bye
January 15, 2015 at 2:55 am
The backups.
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
January 15, 2015 at 4:06 am
I'm with Gail. You find out what the companies expectations are around Recovery Point Objectives (RPO) and Recovery Time Objectives (RTO). RPO is the amount of information, measured in time, that the company is willing to lose. Most organizations it's between 5 & 15 minutes, depending on the database. Make sure your backups can protect at that level. RTO is the amount of time it's going to take to recover to the RPO. Can you do that with your backup scheme.
Once you have that covered, and it's the most important. Then you look at monitoring all the rest of information. And 1-4 are largely concerned with ensuring that you have adequate and appropriate monitoring in place to ensure you can meet all the other needs of the business. Monitoring is done through perfmon, dynamic management objects and extended events (in 2014), or, a third party tool.
"The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood"
- Theodore Roosevelt
Author of:
SQL Server Execution Plans
SQL Server Query Performance Tuning
January 15, 2015 at 4:44 am
One tool I would recommend running against your instances to quickly highlight a number of key issues is sp_Blitz http://www.brentozar.com/blitz/
This highlights issues such as backups not taken recently, incorrect configuration such as autoshrink or autoclose being turned on (third party suppliers seem to be particularly keen on this type of imbicility) and a lot of other stuff you may want to have a look at
edited for trypo
I'm a DBA.
I'm not paid to solve problems. I'm paid to prevent them.
January 15, 2015 at 6:02 am
IMHO:
1. When are the backups being taken?
2. Where are the backups stored?
3. Are the backups viable? In other words, can they be restored.
4. Is the SA account enabled?
5. Everything else.
January 16, 2015 at 12:54 am
GilaMonster (1/15/2015)
The backups.
Ditto that.
--Jeff Moden
Change is inevitable... Change for the better is not.
January 20, 2015 at 8:09 pm
#1. Backups
- Ensure all db's are being backed up(system plus user dbs)
- Ensure transaction logs backups are implemented where needed (check DB Recovery Models)
- Check whether backups are native backups or 3rd party(ex. Red-Gate SQLBackups or Quest LiteSpeed)
- Ensure to test the db backups on a test server using restores
#2. SQL Jobs / Replication / Reporting Services / Analysis Services / Integration Services
- Check on all and ensure functional
#3. Logs
- Check SQL Server Error Logs, Windows Logs for issues
#4. Security
- Server-Level Security, Services Security, Logins/Users Security, Audit Security, File System Security
#5. Baseline the System
- Use tools like Performance Monitor, SQL Profiler, Active SPID Performance
#6. Implement a Monitoring, Alerting & Performance Tool
- http://www.sqlserverops.com, Microsoft Systems Center, http://www.idera.com, Quest, Red-Gate
#7. Implement a Change Control System for all Production Changes
- all changes go through a process(request ticket, change ticket, approvals, deployments)
Relax, enjoy & then start Documentation(Servers, Versions, Service Packs, SQL Jobs, Services, Processes, Permissions)
🙂
January 21, 2015 at 2:22 am
Thankyou very much
😀
January 21, 2015 at 8:05 am
Again, this is for what the first things are that need to be done. The most important things are the backups and basic security, as in those things need to be done in the first day. Jobs, monitoring and all that other stuff, although very important, is totally secondary.
--Jeff Moden
Change is inevitable... Change for the better is not.
January 22, 2015 at 9:21 pm
THX All
and dba-120883, About your reply.
In step 5 :
#5. Baseline the System
- Use tools like Performance Monitor, SQL Profiler, Active SPID Performance
1- Which Counters Do I control in performance monitor in first step?
- I only Check Life_time_exp now.
2- how to Use sql profiler for this check performance ?
- I only check the queries with duration greater than 1000 ms. now.
3- what is spid performance ?
thank you
January 22, 2015 at 10:01 pm
Actually, I'm totally wrong. The 2 most important things are...
1. Locate the coffee machine.
2. Locate the Men's Room.
😛
--Jeff Moden
Change is inevitable... Change for the better is not.
January 22, 2015 at 10:10 pm
sm_iransoftware (1/22/2015)
THX Alland dba-120883, About your reply.
In step 5 :
#5. Baseline the System
- Use tools like Performance Monitor, SQL Profiler, Active SPID Performance
1- Which Counters Do I control in performance monitor in first step?
- I only Check Life_time_exp now.
[font="Arial Black"]2- how to Use sql profiler for this check performance ?
- I only check the queries with duration greater than 1000 ms. now.
[/font]3- what is spid performance ?
thank you
Shifting gears a bit, the worst problems are usually less than 1,000ms but occur thousands of times per day or, sometimes, per hour. Once you've identified such queries, then you could use profiler to capture the "text" for the query, which will usually include the parameters used so that you can use those parameters for testing and optimization. To find such queries, start by right clicking on the instance in the Object Explorer window, select "Reports" from that, and then follow your nose. Understand that the worst queries can change during the day. If nothing resets the DMVs behind the scenes (usually the result of certain nightly runs), then make sure that you make your judgment based on the last compiled date of the query.
--Jeff Moden
Change is inevitable... Change for the better is not.
January 23, 2015 at 1:38 am
sm_iransoftware (1/22/2015)
- I only check the queries with duration greater than 1000 ms. now.
Why?
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
January 23, 2015 at 2:11 am
Because there are bad query.
then I try to optimize them. (But I say this is hard work)
January 23, 2015 at 2:22 am
So a query that runs once an hour and takes 1001 ms is a bad query, but one that runs 50 times a minute and takes 900 ms is a good query?
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
Viewing 15 posts - 1 through 15 (of 31 total)
You must be logged in to reply to this topic. Login to reply