01/16/2004 Mssql2

[Maayan]

My e-mail is full with customer complains.

What about DatabaseMentors suggestion from above - for an alternative connection? Please consider this seriously...

- Maayan

[Mike Bell]

My DB has "suspect" beside it in EM. Nothing is in it. What's that one about?

[DatabaseMentors]

This is a follow-up to my earlier post about the more "fail-safe" approach using syncronized databases across various servers.

I took a little time to code a VB.NET approach of setting the connections strings in the APPLICATION_START of GLOBAL.ASAX using the Application Variables as arrays. And I have drafted 2 functions, which should be stored in a vb file included by each page that opens connections. The first function is GetConnectionString that will return the string to be set in SqlClient.SqlConnection.ConnectionString, but before using it, make sure it is not empty:
cnStr = GetConnectionString
if cnStr.length = 0 then
'no good connections
Server.transfer("PROBLEM.ASPX")
end if
SqlConnection.ConnectionString = cn

The other function is ReplaceFailedConnection(ConnectionString) which will replace the connection string with a good connection (or empty string) and mark the previous as a failure. This is an example of its usage:
Try
SqlConnection.Open
Catch ex as exception
ReplaceFailedConnection(cnStr)
'Note that a new attempt should be made (assuming cnStr is not empty)
End Try

The functions are in the attached text file.

DISCLAIMER: I have not tested the code, and to not represent it to work properly. This was meant as a sample of the logic which could be used to rotate connection strings marking failed attempts.

Hope this will be helpful for some others here... the limitation is to have the servers set up with syncronized databases.

Scott Friend
Database Mentors

[vawebdev]

Would someone from Vortech please give us an update? It has been an hour and 15 minutes since the server was supposed to be back up, but obviously there are a few of us havinging issues still. Thanks in advance.

We've restored the majority of databases so far and they should be working. We (that means Brad) are trying to bring up the remainder--some that could not be automatically restored--as soon as possible and have Microsoft looking at the problem too.

[hostpsi]

I know you guys are trying but should you not really have an extra SQL server on standby for this type of problem? My customers do not understand or accept downtimes that last this long. Please help!

[Brasil]

Are we able to create new DBs by now?

(I am asking cause we don´t wanna keep poking the server if it´s not ready...)

Thanks

[admin]

I have MS on the phone, some of the databases are being marked as suspect and not sure why. The only DB's not working right now are thoses. Thats about 30% or so.

[hostpsi]

My database is still not working (portal-3947844). Can you give me an estimate?

Thanks!

[Maayan]

one of these 30% is my 100% database!

[Brasil]

Originally Posted by admin

I have MS on the phone, some of the databases are being marked as suspect and not sure why. The only DB's not working right now are thoses. Thats about 30% or so.

One of our clients is not being able to create a db. It says that it already exists, but it does not show on the cp.

We put a ticket about it, ticket number [CRV-22503], a few hours back when the problem started.

Would you look at this ticket for us please? And see if you could create it?

Thanks very much

[vawebdev]

Originally Posted by admin

I have MS on the phone, some of the databases are being marked as suspect and not sure why. The only DB's not working right now are thoses. Thats about 30% or so.

I just did a quick count in EM and there appear to be 182 databases marked as "suspect" out of 381 total. For anyone keeping score that is 47.8%.

Is there an estimate or is MS giving any sort of explanation?

Microsoft advised us to reinstall the OS so that's what we'll be doing shortly. We're looking at several hours--as long as it takes to install 2003 Server, SQL server, a service pack and some paraphernalia. Brad and Matt are on their way to the datacentre.

[Brangwyn]

Have you run a low level sector check on the drives at all ?

Able to briefly document the symptoms of the problem (I might have seen it before and can offer some other suggestions)

[admin]

Brangwyn, it was an IBM drive in the system, since they seem to be known to have bad drives not. I am just going to replace it. I have the OS installed now doing MS updates then will do mssql on next reboot and mssql SP3 and get the old drive back in there. We copy all the good DB's over so they will be right up todate. The others will have to come from last nights backup. But since it killed its self around 10am the backup did not run till 11PM the night before the data should not be to far behind.