Matt Dawdy's questions

I've got a server hosted at Rackspace that is a pretty beefy machine (dual hex core, 48 GB of RAM) and our database isn't all that big (100GB or so). SQL Server 2008 R2 Standard 64 bit.

Normally, this machine runs perfectly. Very little pegging of any CPU cores, query response times are great, etc.

However, I can kill this machine's response times by running a query of "SELECT * FROM RawLead WHERE emailAddress = '[email protected]' ; " EDIT: And no, emailAddress is not indexed. Yes, I know that would help this specific query, but the original problem would still remain -- how 1 query can bring the entire server to a halt.

When I do this, SQL Server stops responding to any other queries until this result set is returned.

I just tried this, and as you can see from this screen capture from the Resource Monitor Disk activity changed, CPU didn't really, but the network traffic dropped to basically zero.

Is there a setting in SQL Server that would affect this?

Long story short -- we have 2 webservers (IIS 6) that run a 3rd party asp.net application. Randomly (so far) they just...stop working. I have an outside check that will tell me when it stops working within a minute or so. Right now I have to get onto the machines through RDP and issue an iisreset. Which is fine until I'm not at a machine and I have to get to one PDQ.

I wrote a simple page that will issue an iisreset on the offending remote machine(s). This works, usually. Sometimes "iisrestart \machinename" will stop the IIS service, but not restart it, which is bad.

Ideally, I'd like to know if I can just stop the service, try to start it, and if it doesn't start in 10 seconds, try to start it again. But I don't know how to monitor the status of a service remotely.

Can someone point me in the right direction?

I'm sure this is going to be very dependent on the harddrives and other hardware that we have in our machine, but I have no idea how to read these graphs.

This graph looks nice and low.

This one looks scary as hell.

I know that the scale is different, but how do I know when things are starting to go wrong as far as the disk actually struggling to keep up?

I've got a table with 9 million rows (just received it from our underwriting provider). 4 columns -- leadid (guid), RawLeadXML (nvarchar(max) that contain an xml doc in each row that is 2-3kb), LeadStatusID (int), and id (autonumber int, I just added this).

I am unable to run even a simple delete statement.

DELETE FROM Leads WHERE LeadID = '100a7927-5311-4f12-8fe3-95c079d32dd4' 

I have tried to add an index to LeadID, but that times out.

CREATE NONCLUSTERED INDEX IX_Leads ON dbo.Leads(LeadID) WITH( STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY] 

My machine has 4 processors and 12 GB of RAM. Running Windows 7. SQL Server 2008 (I believe Developer Edition). How can I add an index to this table?

I've got a backup of my SQL Server 2005 database that I want to put on my dev machine (much quicker to develop against than via writing queries through remote desktop.)

I was able to restore this on one machine at work, so I know that the backup is good (well, I assume it is -- I did have to copy it to an external harddrive and then copy it off again onto my current machine).

The error I'm getting is this:

TITLE: Microsoft SQL Server Management Studio

---

Restore failed for Server 'XPS17200911\SQL2005'. (Microsoft.SqlServer.SmoExtended)

For help, click: http://go.microsoft.com/fwlink?ProdName=Microsoft+SQL+Server&ProdVer=10.50.1447.4+((KJ_RTM).100213-0103+)&EvtSrc=Microsoft.SqlServer.Management.Smo.ExceptionTemplates.FailedOperationExceptionText&EvtID=Restore+Server&LinkId=20476

---

ADDITIONAL INFORMATION:

System.Data.SqlClient.SqlError: File 'C:\Program Files (x86)\Microsoft SQL Server\MSSQL.2\MSSQL\DATA\TDC.ndf' is claimed by 'BankA_Tables02'(4) and 'BankA_Tables01'(3). The WITH MOVE clause can be used to relocate one or more files. (Microsoft.SqlServer.Smo)

For help, click: http://go.microsoft.com/fwlink?ProdName=Microsoft+SQL+Server&ProdVer=10.50.1447.4+((KJ_RTM).100213-0103+)&LinkId=20476

And the SQL Script that is generated (you know what I mean) is this:

RESTORE DATABASE [TDC] FROM  DISK = N'C:\Users\Matt\Desktop\banka.bak' WITH  FILE = 1,  
MOVE N'BankA' TO N'C:\Program Files (x86)\Microsoft SQL Server\MSSQL.2\MSSQL\DATA\TDC.mdf',  
MOVE N'BankA_Tables01' TO N'C:\Program Files (x86)\Microsoft SQL Server\MSSQL.2\MSSQL\DATA\TDC.ndf',  
MOVE N'BankA_Tables02' TO N'C:\Program Files (x86)\Microsoft SQL Server\MSSQL.2\MSSQL\DATA\TDC.ndf',  
MOVE N'BankA_Tables03' TO N'C:\Program Files (x86)\Microsoft SQL Server\MSSQL.2\MSSQL\DATA\TDC.ndf',  
MOVE N'BankA_Log' TO N'C:\Program Files (x86)\Microsoft SQL Server\MSSQL.2\MSSQL\DATA\TDC.ldf',  
NOUNLOAD,  REPLACE,  STATS = 10
GO

I formatted it a little bit, but only with carriage returns.

I tried to Google this sucker, but the specific file names might be fouling it up. Plus all of the results seemed to assume a simple database format (this one has 5 files, not just 2 like "normal" SQL Server databases) and all dealt with changing filenames, and I don't think that's my problem here.

Any help would be greatly appreciated.

After a complete failure of an IT person came to my client's office to upgrade the harddrives on a server (windows 2003 RC2), our server now is basically unusable. The machine will boot up but after about an hour of uptime, something happens, not sure what, and the 2nd CPU goes completely red in Task Manager. Kernel mode and CPU usage are at 100%.

The System event log gets filled with this error -- 4-5 a minute:

The driver for device \Device\Scsi\viamraid1 detected a port timeout due to prolonged inactivity. All associated busses were reset in an effort to clear the condition.

I even let this run for 9 hours today after I left and there were 2000+ of those messages in that time frame.

The server is unusable, and has rendered my client completely unable to do business. I'm not an IT guy (I'm a programmer), it's Thanksgiving, and I am completely out of my element.

Anyone have any ideas about that message? Ever seen it before? Ever solved it?

More info: The server has 2 drives in a RAID 0 (I think, that or RAID 1) array. SCSI drives. The previous IT guy got it so messed up that he took the drives out of RAID supposedly, and now when booting, it has to boot to drive 0 just enough to read boot.ini, then we have to choose to boot from drive 1. We can't just boot to drive 1 for some reason.

At first I thought the issue was SQL Server related, as right before the server flipped out we had started an intensive query, but even after I stopped all SQL services, rebooted, etc, it still flipped out on its own after an hour. NOTHING was happening on the server. I mean, no one is in the office, no processes were started (that I know of), etc. Just flipped out.

I've got a table with a primary key that is an int, auto-number auto-increment. Your standard primary key type field.

Due to a complicated scenario with a previously replicated database, now I've got rows already in the table with values up to 410000 yet the value report from IDENT_CURRENT is only 400060 or so. Now, I've got code that is trying to insert records and blowing up because the database thinks that it should use number 400060, and yet I already have a record with that id. Kaboom.

I've looked into manually setting the next number in the table designer, but when viewing the script it generates, it is going to create a duplicate table, copy all records to it, drop the old table, rename it, do some stuff with constraints, etc. It's all very scary when I just need a simple change.

Is there a back way to change this value that won't cause so many unneeded changes?

Okay, I'm not a server administrator, a network guy, or a DBA. I'm merely a programmer helping out a small company. They have IT guy who isn't MS centric (most stuff is on Mac) and he and I are trying to figure out a solution here.

We've got 1 main database. We run nightly full backups. I know they are full backups because I can take the latest file, or any of the daily backups, and go to a completely new machine and "restore" the backup to an empty database and our app runs perfectly fine off of this backup.

The backups have grown from 60 MB to 250MB over 4 months. When running, then log file is 1.7 GB, and the data file is only 200-300 MB. Yes, recovery model is set to full.

So, my question, after all of that, if we are keeping daily backups, and we don't have the need / aren't smart enough to roll the DB back to a certain time, if I change the recovery mode to simple, am I really losing anything?

And, if I do change it to simple, will it completely dump the log file or at least reduce it way the hell down?

And, will that make our database run faster? I know that it'll make my life easier when I copy a relatively recent backup to my local machine to do development and testing...

"The identity range managed by replication is full and must be updated by a replication agent."

This has started happening. We recently broke replication on purpose and don't plan to re-establish the replication.

What else do I need to do to fix this error? Does anyone have any advice for me?

Long story short is I need to get an SMTP server setup inside a client's network that I can send emails out of. This thing won't need to receive emails ever, and really only needs 1 email account customercare.

What can I do? Will I have to have the networking company open a port in the firewall? Is there one built into Microsoft Server 2003?

Long story: client paid networking company to move their website in house. Now I have to figure out how to email out from their website even though they don't have an SMTP server. At least until I install one.

Their email is hosted with Gmail right now (the client's domain through Google App for Your Domain). I changed my code to connect as one of their users "[email protected]" and send email. Worked great for about 12 hours.

All of a sudden none of the automated emails are going out now, and google is sending the emails back saying that it is a permanent failure and Message Rejected. The link they direct me to, http://mail.google.com/support/bin/answer.py?answer=69585 is telling me that our emails look like spam. They aren't. They are emails we send to out clients about the status of their applications. Seriously, they are NOT spam.

What started today as an inability to get to the internet (but people could get in just fine), morphed to we realized that the DNS Server wasn't working, then we figured out that we had a trojan called DNSHost.exe (spybot.rl I think), and we disabled its service entry and deleted the offending file and all registry keys told to use by the Trend Micro site.

Now, we can get on the internet, but the printer being served by this machine (called server2) cannot be printed to from any client machine on the network.

We get the error "The RPC Server is unavailable".

I'm assuming that this is related to the DNS issue we had earlier, as we were able to print just fine until this fun happiness started this morning.

Anyone have any solid suggestions? Windows Server 2003 R2 SP2, and the client machine are all Windows XP SP2.

This is a continuation of a question that I already posted, but this is distinct enough to warrant a new question I believe.

I've got a client who knows very little about how their domains are managed and hosted. After some digging, I found out that the domain is registered through Tucows, but that another company (GSI) is the Registrar of record.

GSI is also the hosting company. Now, when I login to GSI's control panel, I can view DNS information and change it. One of the options is "Start of Authority Server", and its value is "dns.greensoft.com."

So, my question...if I wanted to move this completely away from GSI (they are charging outrageous hosting fees of $800 per month for a site that does 10 MB of traffic. You read that right...), then I need to do 2 things. First, move the registrar from Tucows/GSI to Godaddy, for instance...what the heck do I do? Is the "Start of Authority Server" the setting I need to change to Godaddy?

I have other domains, and the Registrar (Godaddy, Network Solutions) is where I go to set the Nameservers. The nameservers are at the hosting company that I eventually use (if I'm remembering this right) and from there I can setup MX, A, CNAME etc records.

Any help would be appreciated to let me know what I need to do to get my client away from this ridiculous hosting deal.

I've done a Whois and see that the domain I need to work with is registered to Tucows. A company called GSI is actually hosting it.

My client, who owns the domain but knows basically nothing about this, wants to change hosting providers (legal requirement).

So, in order to do this, I will have to somehow get into Tucows, right? And change the name servers from the current hosting company to the new nameservers we want to use?

And, the new company who would be hosting the site, from their control panel is where I would go to set things like MX records (email is hosted by google apps for your domain, so those won't change...but I'm assuming I'll need to set them up again.)

Or am I missing something, and all of this should be done at Tucows?

Right now, I can't get into Tucows. The client has no idea what email/username/password was used to register the domain.

On the current local hosting company's control panel, I have in the past used this to edit things like MX records.

So, I can set up the MX records at Tucows or the hosting company? With Tucows probably winning out if it was done at both places?

My head hurts. I hope this makes sense. Thanks for wading through my mass of probably dumb questions.

EDIT: The site is being hosted at GSI as well.

I assume that there is something at Tucows that exists that tells computers requesting the domain name that they should look to GSI to find information about that domain.

So, per one answer below, what I need to do is this:

1, get into Tucows and see what information is there. Probably enter in some nameserver information that points to the new hosting company.

2, go to the new hosting company and create all the A, MX, etc records that I need to do to support the new setup.

I've got a friend who has been completely abandoned by his previous website guy. Yes, the guy was cut rate, and yes, you get what you pay for.

Anyway, he has asked me to help him untangle all that has been done. Starting now, he wants to know who is actually hosting his website. He knows for sure that the site is registered through GoDaddy, and he previously had the domain hosted there as well. He even prepaid his hosting contract through August 2011 I believe.

Now he's receiving hosting bills from 2 other companies as well (I don't know the names of the other 2 companies at this time). He THINKS that the programmer (who has dropped off the face of the earth) had moved it somewhere, but he doesn't know where it was moved.

Is there any way to tell from a WhoIs or some other method where the site is being hosted? Whois reports 1 IP address, traceroute reports a different IP address.

I'm more than a little outside my comfort zone. Any help or pointers would be much appreciated.

I asked this question on stackoverflow.com, and they recommended that I ask it here, too.

I've got a new client, blah blah blah, I need to break SQL Server 2000 replication from their home office to another database that they are going to completely drop (at some hosting place, it is no longer needed).

The one at the home office is the "Publisher", I believe. Honestly, I think we could just delete the other database and the home office one would run fine, but it would be nice to clean it up a bit and get rid of the publications/subscriptions stuff.

The one issues I'm concerned about is the identity insert. On tables with an auto-assigned integer primary key, each database would "reserve" a block of ids that it would use so they wouldn't trample on each other when synching up.

So, if I just dump the replication stuff, will the database insert happily until it hits the upper bound of its auto-number range? What happens then?

I've been using an iMac for a while at work (1.5 years), and it is getting flaky, so we are replacing it with a new one. Is there a way I can copy all my Mac stuff AND my complete Windows bootcamp partition to the new mac with a minimum of fuss?

I think the subject says it all, but I couldn't find any info quite fitting my desire anywhere. It's hard to search for this and not get tons of articles involving hackintoshes or running a VM on mac.

I want to boot natively into Windows XP (that's working just great).

Currently, I can boot into Mac, too, and run XP in a VM, but I really don't want to do it that way as I spend most of my time in Windows.

So, is there a way to, on the Windows side, buy another version of VMWare fusion (or whatever I need), create a new vhd, install mac osx on it, and make my life about 10x easier than it currently is?

Thanks for any and all help.