Production Server Down - help!

Hi All,
Newbie here, and in typical newbie fashion, has managed to down our
"live" server. People are back on Monday, so I have this wend to fix
the problem.
Was running an operation on the DB via my ERP s/w (deleting a user
created field) which was taking far longer than normal. Being impatient
(and thinking it had crashed), I killed the app. Restarting the app
failed, so I decided to take a look at the server.
EM first of all looked like it had died, so stopped the SQL service.
Have looked at the event logs and it seems that :
3450:
Recovery of database 'XYZ' (9) is 0% complete (approximately 10011 more
seconds) (Phase 2 of 3).
Now, I am happy to start the SQL Server service, and EM, and literally
sit and wait as it were. My main concern is that the sqlserver process
was eating up huge amounts of memory. Won't my system just keel over
after a bit? The DB is 9gig and the log file is 9gig (normally max of
say 300mb) - is this normal? Given I was only deleting a single field
in the DB, what are the chances I have corrupted any other data? My
last backup is from Thursday PM.
This learning experience is going to stick...
Thanks
RajivWhat you did by restarting SQL Server is force a rollback to start. I'm
betting that EM shows your database as in recovery?
Get yourself a drink, find a comfortable chair, and wait. It might take a
while, but it will eventually come back online.
Adam Machanic
Pro SQL Server 2005, available now
http://www.apress.com/book/bookDisplay.html?bID=457
--
"abd08" <dejaonly@.hotmail.com> wrote in message
news:1148079787.604179.235610@.j33g2000cwa.googlegroups.com...
> Hi All,
> Newbie here, and in typical newbie fashion, has managed to down our
> "live" server. People are back on Monday, so I have this wend to fix
> the problem.
> Was running an operation on the DB via my ERP s/w (deleting a user
> created field) which was taking far longer than normal. Being impatient
> (and thinking it had crashed), I killed the app. Restarting the app
> failed, so I decided to take a look at the server.
> EM first of all looked like it had died, so stopped the SQL service.
> Have looked at the event logs and it seems that :
> 3450:
> Recovery of database 'XYZ' (9) is 0% complete (approximately 10011 more
> seconds) (Phase 2 of 3).
> Now, I am happy to start the SQL Server service, and EM, and literally
> sit and wait as it were. My main concern is that the sqlserver process
> was eating up huge amounts of memory. Won't my system just keel over
> after a bit? The DB is 9gig and the log file is 9gig (normally max of
> say 300mb) - is this normal? Given I was only deleting a single field
> in the DB, what are the chances I have corrupted any other data? My
> last backup is from Thursday PM.
> This learning experience is going to stick...
> Thanks
> Rajiv
>|||Ahhh, I remember the first time I killed a server. 30-something hours
over a stupid mistake; luckily I was too wet behind the ears to play
with a production box.
As Adam suggested, get a drink and a chair and wait it out. It'll
clear out eventually; however, you do need to take a look at your log
files and try to understand what caused the delete process to hang in
the first place.
Oh, and welcome to the club. You'll make a fine DBA someday.
Stu|||I'll chime in here since nobody answered your memory question. SQL will
consume nearly all your system memory. It will stop when there is around
8-10 MB free. This is by design.
Oh, and yes, I too have done the "hard kill" on a server and the "long wait"
while it recovers. It is one of the rites of passage. Slow learners do it
more than once.
Geoff N. Hiten
Senior Database Administrator
Microsoft SQL Server MVP
"abd08" <dejaonly@.hotmail.com> wrote in message
news:1148079787.604179.235610@.j33g2000cwa.googlegroups.com...
> Hi All,
> Newbie here, and in typical newbie fashion, has managed to down our
> "live" server. People are back on Monday, so I have this wend to fix
> the problem.
> Was running an operation on the DB via my ERP s/w (deleting a user
> created field) which was taking far longer than normal. Being impatient
> (and thinking it had crashed), I killed the app. Restarting the app
> failed, so I decided to take a look at the server.
> EM first of all looked like it had died, so stopped the SQL service.
> Have looked at the event logs and it seems that :
> 3450:
> Recovery of database 'XYZ' (9) is 0% complete (approximately 10011 more
> seconds) (Phase 2 of 3).
> Now, I am happy to start the SQL Server service, and EM, and literally
> sit and wait as it were. My main concern is that the sqlserver process
> was eating up huge amounts of memory. Won't my system just keel over
> after a bit? The DB is 9gig and the log file is 9gig (normally max of
> say 300mb) - is this normal? Given I was only deleting a single field
> in the DB, what are the chances I have corrupted any other data? My
> last backup is from Thursday PM.
> This learning experience is going to stick...
> Thanks
> Rajiv
>|||"Geoff N. Hiten" <SQLCraftsman@.gmail.com> wrote in message
news:%23ADuSe6eGHA.3364@.TK2MSFTNGP05.phx.gbl...
> while it recovers. It is one of the rites of passage. Slow learners do it
> more than once.
*guilty whistle* ... Luckily I never did it on a production box, but I
did once bring an entire dev team pretty much to a halt for two working days
near the end of a release cycle... gained many friends amongst the
developers for giving them a bit of wiggle room, and was never totally
forgiven by management :)
Adam Machanic
Pro SQL Server 2005, available now
http://www.apress.com/book/bookDisplay.html?bID=457
--|||Thanks for the support/info guys!! SQL is now in the "recovery"
process, so will log back into the server tonight/tm am and see if
everything is intact. Will also have a long look over the log files and
see what could have caused it.
Always good to know though that am not the only one who has killed a
server in this way!! =)
Thanks again... still lots to learn!!!!
Rajiv