MikeMurko's questions

Our backup procedure is as follows: midnight full backup, every 8 hour differential backups, and every 5 minutes transaction log backups. Maintenance plans keep everything rolling and tidy.

To restore to a specific point in time (let's say yesterday at 2:05pm). On restoring with SQL Server Management Studio 2012, we right-click the database we want to restore to > Tasks > Restore > Database. We then restore the full backup with no recovery (and overwrite existing data). This leaves the database in a "Restoring..." state. We then go to > Tasks > Restore > Files and Filegroups. Select our closest diff backup (noon). Restore with no recovery.

Now for the transaction logs - Tasks > Restore > Transaction Log. On that dialog we select "From file or tape", and click the button to choose files. Add, and select the 5 minute transaction logs from 12:00PM to 14:05PM (25 of them). When we click "OK" on the dialog box ... SSMS crashes and gives the following error:

An exception occurred while executing a Transact-SQL statement or batch. (Microsoft.SqlServer.ConnectionInfo)

The media loaded on "C:...\backup_2012_12_22_120500_4174134.trn" is formatted to support 1 media families, but 25 media families are expected according to the backup device specification. RESTORE HEADERONLY is terminating abnormally. (Microsoft SQL Server, Error: 3231)

For help, click: http://go.microsoft.com/fwlink?ProdName=Microsoft%20SQL%20Server&ProdVer=11.00.2218&EvtSrc=MSSQLServer&EvtID=3231&LinkId=20476

If I recall this is the exact procedure we used in SQL Server 2008 and 2005, so I cannot see why this is failing. Is this something specific to 2012? Is it a bug? I couldn't find any information on this online. We are not using tape backups at all and most of the stuff I read about media families has to do with tape backups.

Troubleshooting: Selecting each transaction log (one at a time) in sequence and restoring works. However this can easily take 15+ minutes with so many transaction logs.

I haven't tried, but I think if we wrote plain old TSQL in the form:

RESTORE DATABASE [OurDB] FILE = N'db_dat' FROM  DISK = N'C:\...\diff\backup_2012_12_23_085000_4627006.bak' WITH  FILE = 1,  NORECOVERY,  NOUNLOAD,  STATS = 10

And ran this for each xaction log - it should work. I can write some powershell or something to restore the database ... but shouldn't the "Management Studio" be able to handle this? This answer seems to point to "No". https://dba.stackexchange.com/questions/1021/how-to-restore-multiple-backups ... however the GUI appears to hint at allowing multiple restorations "Specify the source and location of the transaction log backups" - not to mention it lets you select multiple files in the first place.

I would like to use Amazon EC2 as an emergency backup database+web server in the event our primary host becomes unavailable. I feel like I wouldn't have trouble setting up a Windows instance, install SQL Server and get the web server up and running (would take a few hours, plus installing various libraries, our source code, etc).

My question relates to pricing. If I simply "stop" the instance rather than "terminate" it, does that stop counting "instance-hours"? I would prefer not to terminate the instance and lose all that work I spent setting it up. If I must "terminate" in order to stop the billing - is it possible to make an image of the server after I have set it all up, then save that image somewhere (S3?) Is this something that people do regularly?

Ideally this instance would just be waiting in the wings for an issue with our host, but costing us nothing except perhaps data storage costs.

After talking to a friend who works for a virtualization firm, I started to think whether it could be of benefit to my small company. Right now we have our IIS web server on one server, and our SQL database on another. What he was saying is that with virtualization you could put both operating systems and application stacks onto one piece of physical hardware and then just get a massive big server with lots of CPU, RAM, and RAID 5 or 10 for redundancy. This saves us on the costs of buying essentially duplicate amounts of hardware to backup each IIS and database server (I think?)

This all sounds fine and dandy, but what is the difference compared to me simply installing IIS onto my SQL Server and serving websites/data on the same box with 1 operating system? To me I save the overhead of having a virtualization environment, but achieve the same goal. Also, are there not performance concerns with having all your eggs in one server-basket?

One disadvantage that I could see is that it would be more difficult to split those two "logical" servers at some point down the road. However, I feel like I am missing the point of virtualization....

We have set up all the equipment necessary to have a gigabit network. All servers + workstations are running gigabit cards, just bought a 16 port switch that can handle gigabit as well. My concern is now that our firewall is not up to scratch. Our firewall is a Juniper Netscreen 5GT which is only 10/100 mbit. If all of our computers are set up to be inside the "trusted zone" - can I still realize the gigabit potential without upgrading our firewall. My thinking is that all computers will talk to eachother through the switch and shouldn't run into any trouble with the firewall - however I have read that a single 10/100 device brings down the whole network to a slower speed even if two computers talking are gigabit.

If not, does anyone recommend the next highest model in the Juniper line that can handle a gigabit output (all I would need is one port as our switch can handle the rest).

Here is a handy diagram of our future setup. Green are all gigabit devices.

We have a RAID 5 array (3 drives) one of which failed this morning. If we take that drive out the system can boot into Windows Server '03 fine. Problem being we don't have anywhere to go if another drive fails. We have a good drive which we plugged into the failed one's slot while the server was in Intel RAID BIOS (not WIndows). Hit the "rebuild" button and it was slowly crunching.

We need this server to be back up and running so we booted into Windows and using the web-based manager we see that the drive is "rebuilding" .... but we're just not sure whether it actually is. It's been stuck at 1% for about half an hour. 99/250 gbs are used. Is this normal? Is the drive actually rebuilding while Windows is running?

Love to hear from someone who knows what they're doing!