What are you supporting at the moment?

[roobaron]

I noticed in a couple of threads that people find it interesting what other DBAs/Developers are doing in terms of support etc.

I thought a series of questions like
1) Type of databases supported, version and OS
2) Largest database
3) Largest table
4) Types of indexes used
5) Data processing ETL, transactions etc
6) Major tuning issue
7) Highest database wait
8) Major project

So to kick off the thread here is what i support

1) I support both DSS and also DWH databases. Mostly Oracle 8.0.5 running on Tru64 Unix, HP-UX, Solaris. Also support SQLserver 6.5,7,2000.
2) The largest database we have is around 40Gig in total
3) the largest table has 38 Million rows.
4) Combination of normal b*tree and bitmap indexes
5) Most processing involves massaging the data we get and loading/updating the database with this new information.
There is also database extracts to inhouse datamart
6) Poor program logic i.e visiting every row in the large tables, but the program logic means that Oracle chooses an index scan (unique) to do this. The disk I/O maxes out of the no of small I/Os not the I/O throughput.
7) Index scans (db sequential read)
8) Upgrade to 8.1.7

[pando]

Well I dont look after at any specific database because in the shop I am now there are quite a few databases, around 160 instances and the DBAs are divided into 2 groups, engineering and production, production are like slaves I would say, basically do routine jobs, runing scripts, backup & recovery, etc Engineering are those who do perfomance tuning (but do they do it at all?), investigate solutions, evaluate products etc I am not in any of two groups because I am in DWH first line support, pretty boring job but since I am one of the two guys who knows a bit about dba in support sometimes peeps call us two the casual dbas

But anyway in those 160 about 1/4 is production the others are development, prer-production, integration etc. They are OLTP and DWH, since it's a Telco the OLTP has huge amount of connections from 400 to 6000 (the CRM application, billing system), size from 40gb to 120gb. The DWH are bigger from 400GB to 2TB, versions from 7.3.4 to 8.1.7, platform all unix, Tru64 mainly, then hp-ux and solaris

[ratcheer]

I work in the finance dept of a medium sized bank. I support a financial reporting application. My main database has three instances, each about 120 GB. Running 8.0.6.2 on Sun Solaris 2.6 and Sun E-6000 hardware.

I also support two Oracle Express instances on smaller Sun E-3000 boxes.

My biggest issue is the AWFUL SQL that the application software generates. It is about 95% ad hoc, so there is nothing I can do to improve the SQL. All I can do is create indexes and hope they will be used, and throw more hardware/money at the platforms. It is very frustrating.

Especially since at my previous job, our application software was about 95% in-house development and, when something ran poorly, I could coordinate with the developers to make the database and the software work better together. But, in my current job, the software is supplied by a vendor and I have close to zero influence.