How to determine the last access date of a DB?

(SQL Server 2000, SP3a)
Hello all!
In my company, we have several development servers that, over time, get clut
tered with a
*lot* of "scratch" databases. Is there some inherent mechanism that will sh
ow the last
time a database was accessed? If not, is there a relatively simple SQL-base
d solution
that I could craft that could help track that?
Thanks for any help you can provide!
John PetersonYou have to do some server level settings for that.
SQL Server can log event information for logon attempts and you can view it
by reviewing the errorlog. By turning on the
auditing level of SQL Server.
follow these steps to enable auditing of all/successfull connections with
Enterprise Manager in SQL Server:
Expand a server group.
Right-click a server, and then click Properties.
On the Security tab, under Audit Level, click all/success etc(required
option).
You must stop and restart the server for this setting to take effect.
Vishal Parkar
vgparkar@.yahoo.co.in | vgparkar@.hotmail.com|||Login attempts won't tell you if a database has been used or
not. It will only tell you about server logins, not database
access. You would need to monitor this with profiler
going forward. I don't think there is any direct way to get
historical usage if you haven't been monitoring for such.
I suppose you could always drop the databases and then see
who calls complaining to determine if the databases are
being used. And you could use whatever criteria in
determining if you actually backup the databases before
dropping them. That could make for an interesting work day.
-Sue
On Thu, 8 Jul 2004 07:04:50 +0530, "Vishal Parkar"
<REMOVE_THIS_vgparkar@.yahoo.co.in> wrote:

>You have to do some server level settings for that.
>SQL Server can log event information for logon attempts and you can view it
>by reviewing the errorlog. By turning on the
>auditing level of SQL Server.
>follow these steps to enable auditing of all/successfull connections with
>Enterprise Manager in SQL Server:
>Expand a server group.
>Right-click a server, and then click Properties.
>On the Security tab, under Audit Level, click all/success etc(required
>option).
>You must stop and restart the server for this setting to take effect.|||Thanks guys! Yeah...if I could put a trigger or something on sysprocesses,
I could
potentially log access to specific databases. I suppose I could have a "mon
itor" wake up
every X seconds and read the distinct databases that were in use -- but that
seems kind of
heavy handed...
"Sue Hoegemeier" <Sue_H@.nomail.please> wrote in message
news:4kdpe010faeumj5c62626q6oo83iegva7m@.
4ax.com...
> Login attempts won't tell you if a database has been used or
> not. It will only tell you about server logins, not database
> access. You would need to monitor this with profiler
> going forward. I don't think there is any direct way to get
> historical usage if you haven't been monitoring for such.
> I suppose you could always drop the databases and then see
> who calls complaining to determine if the databases are
> being used. And you could use whatever criteria in
> determining if you actually backup the databases before
> dropping them. That could make for an interesting work day.
> -Sue
> On Thu, 8 Jul 2004 07:04:50 +0530, "Vishal Parkar"
> <REMOVE_THIS_vgparkar@.yahoo.co.in> wrote:
>
>|||Triggers on system tables isn't supported though and
sysprocesses is only materialized when used so there
wouldn't even be something to put a trigger on.
-Sue
On Wed, 7 Jul 2004 21:52:56 -0700, "John Peterson"
<j0hnp@.comcast.net> wrote:

>Thanks guys! Yeah...if I could put a trigger or something on sysprocesses,
I could
>potentially log access to specific databases. I suppose I could have a "mo
nitor" wake up
>every X seconds and read the distinct databases that were in use -- but tha
t seems kind of
>heavy handed...
>
>"Sue Hoegemeier" <Sue_H@.nomail.please> wrote in message
> news:4kdpe010faeumj5c62626q6oo83iegva7m@.
4ax.com...
>|||Understood. How about reading database logs? Would *that* show any activit
y? How would
I even go about doing that?
"Sue Hoegemeier" <Sue_H@.nomail.please> wrote in message
news:q4eqe0htf3hbgg1hme377t5ro5gpcvs351@.
4ax.com...
> Triggers on system tables isn't supported though and
> sysprocesses is only materialized when used so there
> wouldn't even be something to put a trigger on.
> -Sue
> On Wed, 7 Jul 2004 21:52:56 -0700, "John Peterson"
> <j0hnp@.comcast.net> wrote:
>
up[vbcol=seagreen]
of[vbcol=seagreen]
>|||Well...I noodled on using DBCC LOG to try and see if any of that information
might be
germane, but I don't think so (everything is "reset" on a restart of SQL Ser
ver).
Then I looked at sysobjects.refdate, but that didn't pan out either.
Is there *any* way to determine the date/time that a particular table was la
st affected by
an INSERT/UPDATE/DELETE?
There seems to be a similar question here:
[url]http://www.experts-exchange.com/Databases/Microsoft_SQL_Server/Q_20684794.html[/ur
l]
But I don't want to sign up just to find out I *can't* do this... ;-)
Thanks for any help you can provide! :-)
"John Peterson" <j0hnp@.comcast.net> wrote in message
news:%23Cd7QnPZEHA.2944@.TK2MSFTNGP11.phx.gbl...
> Understood. How about reading database logs? Would *that* show any activity? Ho
w
would
> I even go about doing that?
>
> "Sue Hoegemeier" <Sue_H@.nomail.please> wrote in message
> news:q4eqe0htf3hbgg1hme377t5ro5gpcvs351@.
4ax.com...
wake[vbcol=seagreen]
> up
kind[vbcol=seagreen]
> of
>|||The SQL Server logs...not reliably. Unless the databases
have errors all the time that get logged. In terms of the
database log, I don't think you can't get timestamps using
dbcc loginfo or dbcc log. You can check object creation
dates in the database but that won't necessarily mean the
database isn't being used if no one is creating objects. And
even with the logs, if for some reason, they are only
selecting out of the database, that wouldn't do you any good
either.
If you want to read details of log files, you need to use
something like Log Explorer from lumigent -
www.lumigent.com
-Sue
On Thu, 8 Jul 2004 07:42:36 -0700, "John Peterson"
<j0hnp@.comcast.net> wrote:

>Understood. How about reading database logs? Would *that* show any activi
ty? How would
>I even go about doing that?
>
>"Sue Hoegemeier" <Sue_H@.nomail.please> wrote in message
> news:q4eqe0htf3hbgg1hme377t5ro5gpcvs351@.
4ax.com...
>up
>of
>