April 6, 2012 at 2:57 pm
Lynn Pettis (4/6/2012)
TravisDBA (4/6/2012)
We strictly handle SQL Server support. No Access, no applications, just SQL Server. We have well over 100 instances of SQL Server to administrate spread throughout the state of Iowa.
I totally agree with most of what you say. However, the problem is with that, is the people reporting the issue(error), as well as HelpDesk personnel on duty many times cannot always determine if the issue is database related, network related, or application(even Access) related. In many companies the Help Desk department has a constant turnover of people, so seasoned individuals that have the experience to detect the difference have often moved on or left. So, as a result, many times a DBA will get called for a wild goose chase anyway. Fortunately, in my company the Help desk people have been around for awhile and can usually ferret this out before the call is made. 😀
My experience has been it is ALWAYS a database issue until proven otherwise beyond any reasonable doubt.
I agree. I don't want to give the impression that we only receive pages or calls about SQL Server. But we are third level support; a problem has to pass through two other levels before we are paged. :doze:
Usually the service desk does what they can to help. If necessary it gets passed along to the application analyst on call who is responsible for the specific application. If the application analyst runs into problem then they contact our layer and the analyst are pretty good at notifying the teams that they think would be the most likely to help such as networking, server, application or client device people. :cool:;-) :hehe:
Our service center does a pretty good job of working out who to contact for a problem, which for SQL Server issues is about 95%+ of the time. The other 5% we scratch our head and wonder why it was sent to us.:crazy:
-----------------
Larry
(What color is your database?)
April 6, 2012 at 7:37 pm
I'm not an enterprise DBA, but I support a couple of applications with off-farm SQL instances. SQL has been rock-solid for the duration of my tenure; most middle-of-the-night calls result from application snafus.
I have to give a huge "THIS" to proper development and code promotion procedures. SQL isn't excused. Adopt a strong test and promotion methodology, and stick to it. At my current job, this isn't an option and is mandated by regulation, so you can be sure it's followed without exception.
If your company doesn't have controls and your cowboy developers can run roughshod over the production environment, become a champion for controls and methods. Do it now.
April 9, 2012 at 8:26 am
I am never on call. I made it a condition of employment. Been there, done that, got the t-shirt a long time ago. My work is highly specialized, so I do get called occasionally when an issue is very complicated or a DBA is stuck trying to figure something out. By occasionally I mean a couple of times a year. We have a very proficient technical staff where I work, and there are firm controls in place to prevent "accidental code" from assaulting production.
April 9, 2012 at 10:00 am
If your company doesn't have controls and your cowboy developers can run roughshod over the production environment, become a champion for controls and methods. Do it now.
This makes me remember an interview I once had where the interviewer started out the interview with "All my developers have sysadmin access to the production environment, Do you have a problem with that?" My immediate return question was "Who restores the production database back to point in time if they screw it up. His answer was "Well..... the DBA, of course." That pretty much terminated the interview for me right there. You can champion all you want, but if management doesn't back your controls and methods, you are pretty much left in the ditch, and it is time to just move on. 😀
"Technology is a weird thing. It brings you great gifts with one hand, and it stabs you in the back with the other. ...:-D"
April 9, 2012 at 10:11 am
JP Dakota (4/9/2012)
I am never on call. I made it a condition of employment.
I am glad for you, but in reality, most DBA's would not get the job in the first place if they made this a condition up front. It's just part of the job for most DBA's.:-D
"Technology is a weird thing. It brings you great gifts with one hand, and it stabs you in the back with the other. ...:-D"
April 9, 2012 at 3:43 pm
I'm one of a pair of DBAs - between us we look after 50 SQL Servers and 35 Oracle Databases. We are on call 24/7, alternating one week on one week off.
In the year I've held this post, I've received just one out of hours call for a broken SQL Server, and at least 5 for Oracle!
April 9, 2012 at 4:17 pm
That's a great record! In my experience, it's usually the DBAs who a) don't know what they're doing or b) get lazy and take shortcuts that find themselves getting lots of calls during the off-hours. There are exceptions, but as they say, an ounce of prevention is worth a pound of cure!
April 10, 2012 at 7:00 am
At my organization I am the only fulltime DBA, responsible for over 50 instances and 700+ SQL Server databases. Thus, I'm on call 24x7x365 (that includes vacations). Prior to being the DBA, I was one of the developers and we had no DBA. Since I was fixing most of the database issues when they happened anyway, I asked to be the fulltime DBA. I used to get calls and pages after hours 4-5 times a week. After 4 years of being the DBA, I get actual phone calls about database issues 1-2 times a year, and pages from automated monitoring I setup 4-5 times a month.
The biggest reason for the downturn is getting the developers to use source code control and automating development environment refreshes for them so they're coding in an environment that matches the production environment they're going to deploy to. They catch a lot more of their errors before the code gets to production and they write a lot better code now. Most of my pages now are warnings about impending disk space issues.
April 10, 2012 at 7:19 am
Our other DBA and I alternate who is on call. The DBA "on call" handles incoming issues during the day and watching monitoring. For off-hours issues they're the first one called but since it's not an official on call (ie not paid for it and don't carry pager) we're not required to respond. If whomever is on call doesn't respond then the other person gets called (which rarely happens). It works well since we're both dedicated and respond if we can. For any time spend working a page we get time off to compensate and our boss is very flexible with that.
Since it's rare to get paged (every few weeks if that) what we have works. My only concern is that at some point either both of us won't be available or something will happen off hours that our monitoring utility catches but since we're not officially on call we won't get paged for and result in downtime only because it's not handled before users come in. However, both my boss and his boss are aware of the concern and say it's acceptable, mostly because every department in our hospital has downtime procedures that allow them to operate in the event of any system failure. Which means if anything does happen I won't need to worry as much about defending the fact that it didn't get caught.
April 10, 2012 at 9:11 am
For off-hours issues they're the first one called but since it's not an official on call (ie not paid for it and don't carry pager) we're not required to respond. If whomever is on call doesn't respond then the other person gets called (which rarely happens). It works well since we're both dedicated and respond if we can
I had this exact same situation at Citrix where my partner was a 21 year old kid that slept so hard a hand grenade could not wake him up and I got called as his backup all the time so I was doing on-call constantly. Citrix was very political and the boss liked the kid and made every excuse for him like for example "Well Travis he takes prescription medication at night". Well I immediately said "Well boohoo, I have a couple of beers at night does that let me off the hook when I am on call and fall asleep?" Citrix managment is some of the worst I have ever seen in over 28 years in this industry. It's almost fast-food (Burger King) oriented managment, Just terrible. 😀
"Technology is a weird thing. It brings you great gifts with one hand, and it stabs you in the back with the other. ...:-D"
April 10, 2012 at 9:22 am
TravisDBA (4/10/2012)
For off-hours issues they're the first one called but since it's not an official on call (ie not paid for it and don't carry pager) we're not required to respond. If whomever is on call doesn't respond then the other person gets called (which rarely happens). It works well since we're both dedicated and respond if we can
I had this exact same situation at Citrix where my partner was a 21 year old kid that slept so hard a hand grenade could not wake him up and I got called as his backup all the time so I was doing on-call constantly. Citrix was very political and the boss liked the kid and made every excuse for him like for example "Well Travis he takes prescription medication at night". Well I immediately said "Well boohoo, I have a couple of beers at night does that let me off the hook when I am on call and fall asleep?" Citrix managment is some of the worst I have ever seen in over 28 years in this industry. It's almost fast-food (Burger King) oriented managment, Just terrible. 😀
At a previous employer we had an on-call rotation as well. At one point we had 4 developers, so on-call went three levels deep giving one person a week off from any any on-call responsibility. The only problem we had was one developer. If you followed this person in the rotation you were guaranteed to get the call as this person never answered their phone at night. Oh, and this person never got in trouble for not responding.
December 12, 2013 at 2:06 am
Felt like digging this thread. I'm 1 of 2 DBAs and we're on call on an every other month basis. We haven't had an out-of-hours call now in months that has been an SQL problem.
Our issue is, constant application upgrades/patches. We're supporting over 100+ applications across 30 odd Prod instances, and there's at least 12 of these upgrades needing to be done each month.
And now the boss in IT is trying to suck up to the business by saying all upgrades will be done after 9:30pm at night, rather than previously we could start upgrades at 6pm. Everything we script from the db side to pass through the various environments so its pretty fast to deploy but its a real pain as some take 3/4 hours for the apps guys to do the various app side upgrade and they need us on stand-by the whole time if they need a rollback done.
Viewing 12 posts - 16 through 26 (of 26 total)
You must be logged in to reply to this topic. Login to reply