OnQ

On MSDE

Installing named MSDE instances on a server is a pretty common thing. Or so i thought until I ran into two applications that have given me fits. I mean, it's been a long time since I've been a programmer (well, at least a paid one), but seriously, it's not that hard to figure out how to install a named MSDE instance so that it doesn't conflict with another database already on the box.

Right?

Apparently not. One vendor has an installer that simply dies if any other MSDE databases are installed on the box. The installer assumes that there are no other instances (or will be no other instances) so it only checks to see if some of the key components are installed on the box and then aborts the install if they are. Geez! You've never, EVER, had a client want to install your product on a box where another database lives? And you've been in business how long? And you call this version 7?

Only slightly less excusable is a second vendor who, until recently, played well with others. I've done a couple dozen installs of their product on SBS servers (raise your hand if you know how many MSDE databases SBS installs by default - hint: it's more than one) with no issues, but they updated their installer and now it apparently fails to recognize which MSDE slot to try and install into. They've at least got a workaround, but good grief, what happened to quality control and testing?

So to all you database developers out there, please, please, please, please figure out how to correctly install a named instance of your MSDE database on a server, even if there are no other MSDE instances present. It's very short-sighted to assume that you're producing the only MSDE database in the world and can short-cut the detection process during an installation. Your customers and your partners will be very, very happy.

On Amsterdam: Day 2, Jeff's Disaster Recovery Presentation

The last session of the conference I attended was Jeff's new Disaster Recovery presentation. Jeff let us know right up front that this was not a "best practices" session. Instead, he tackled something a little different.

Disaster recovery and disaster preparations are linked. Disasters don't always happen suddenly - they usually happen slowly and over a long time. Jeff related his experience of relocating his wife's company from New Orleans to Texas in response to Hurricane Katrina. He set up a set of slides for the group to discuss if there was interest in the room. These are the topics that were discussed:

• • Myth - you can have no other domain controllers in an SBS domain. Not correct. You can have multiple DCs in the domain, so long as the SBS server holds all the FSMO roles and there can only be one SBS server permanently installed in the domain.
• • Myth - replacing the SBS server NIC is not allowed. This is a holdover from earlier versions of SBS but is no longer the case. This "feature" was dropped in SBS 2000. You can replace NICs in the box, either through booting in Directory Services Restore Mode or by doing the loopback adapter trick.
• • Myth - You cannot move a drive image directly to new hardware. You can, and in some cases it will actually work. Must have:
  • • compatible HAL
  • • accurate boot.ini, consistent boot device order
  • • boot critical drivers are already installed
  There are also 3rd party imaging products that are making this an easier task.
• • Myth - You cannot restore the system state to new hardware. This actually does work, and is different than restoring a drive image. Domain controllers present challenges for AD restore, Exchange SQL, and Monitoring. A baseline install or same hardware ASR disk to prepare for a system state restore. ASR disks do not work for bare metal restore to different hardware.
• • Myth - You cannot do a "drive slide move" into new server hardware. Same rules apply as with the previous two points.
• • Myth - You can repair a server by reinstalling a service pack. Reinstalling a Service Pack generally will not repair missing files and registry entries for an OS. An in-place upgrade as a repair tool generall will be helpful, and reinstalling application service packs is also generally helpful.
• • Myth - Bare metal restore to "known good condition"? short filename restore breaks cause problems with registry/filepath alignment. 2003 is the only OS that can do this successfully. Cold drive imaging is the only solution for true restore to previously good condition. If you do a restore for anything older than 2003, you really need to look at a drive image restore then a file restore from tape to avoid this problem.
• • Myth - IDR and recovery automation. Independent Disaster Recovery is a generalized name for 3rd party product tools to "boot to restore" or "click to restore" IDR systems can have "fragile" requirements and really, really must be tested. Jeff doesn't like this and doesn't use it himself. IDR really needs to have identical hardware, software, etc., and, again, must be tested.
• • Myth - SBS server is too dirty to migrate, I must wipe it and rebuild. Corruption of AD is really rare and shoudl not be assumed. Group policy can cause a lot of problems if its not done correctly, but if you have a good understanding of group policy, you can work through it (it may just take time).
• • Can you save a domain from a dead DC? "Graveyard swing migration" can pull back not only the server but the entire domain. A recovery server can be replaced by Swing Migration with a clean server, and still preserve the domain. Backup DC and recovery of a dead DC provide similar recovery options.
• • Myth - A large C partition is the best! Please, please, please no! Disaster recovery is faster if you have multiple partitions since you really only have to restore/recover the system partition (provided the other partitions are still OK).
• • Myth - OEM media can only build a new domain. So long as you get OS media from the vendor and not "recovery CDs." Most vendors will give you standard install media. OEM media cannot be used for in-place upgrade repairs of non-OEM installations.
• • Last Known Good to recover from SP or patch update failure. Eh, probably not. "Last Known Good" only restores previous system registry settings, but not other hives or driver files. System state restore is the only option for a genereal roll-back (which doesn't always work). Some 3rd party products might be able to do "delta" rollbacks on drive writes.
• • Myth - You can best do a disaster recovery by having cold hardware on site. Since you can restore to different hardware successfully, this is not really an issue.
• • Myth - SBS FSMO roles are risky. Not exctly. By default, all DCs in a domain have exact copies of the AD database by default. Global Catalog roles are not required to move a complete replica in a single domain model.
• • Myth - SBS is a bad disaster reecovery risk. Jeff doens't think so. Challenge comes in that SBS is a DC, is an Exchange server, etc., and it's tough to pull a machine down and rebuild from scratch. But since we can use drive imaging, system state backup, and application data backup, we can get the server back up fairly quickly.
• • Mystery - product activation does not prevent you from replacing a DC motherboard. You will need to reactivate it, but you can get it going again. But what about OEM restrictions on server OS software? it's possible that a vendor may put a set of restrictions on the replacement of certain components that might require you to purchase a new license.
• • Mystery - An Exchange store mount has failed, and probably has corruption in the data. Not necessarily. There are lots of reasons that an Exchagne store won't mount. Forklifting an Exchange database to an older version of Exchange will generate really ugly errors. Disk permissions on the volume where the databases reside, along with permissions on the MDBDATA folder, can also keep the database from mounting.
• • Mystery - Is there no predictable way to recover Exchange databases? Yes, it's possible that doing a physical repair (hard recovery) could seem to trash data. If there is corruption in the database, the hard recovery is going through and finding data that's good and keeping that, which might look like data has gone missing. Always always always make a backup copy of the databases before you run the hard repair.
• • Mystery - Is it critical to have all the Exchange log files in order to recover the information store intact? No. The log files are really additional disaster recovery information. The reality with SBS servers is that the information in the log files is written to the databases very quickly.
• • Mystery - Backup skips files, does that mean it's not a complete system state backup? No, most of the files it skips are created dynamically or are other files that will be handled in other ways.

On Amsterdam: Day 2 (session 3)

I sat in on David Overton's (the keynote speaker) session titled SBS 2003 - Level 501. This was a troubleshooting/scenario session. Also known as why not to do a demo running a beta of Vista with a beta of Virtual PC to host a beta of SBS 2003 R2.

• • Adding search to WSS using SBS Premium edition (this is based on R2)
  • • Follow the install instructions (sqlinstallsteps.htm on the Premium Technologies CD)
  • • Make sure to install the Full Text Search Engine unless you really, really don't want search capabilities
  • • Using Ifilters to add functionality (such as seaching PDF files using the Adobe Ifilter, if you install Office (but not Outlook) the TIFF Ifiliter will get installed and you can search FAXes if you set up FAXes to get stored in Sharepoint)
• • Adding URL/site blocking/allowing to ISA 2004
  • http://www.isaserver.org/articles/2004firewallblocklist.html
  • http://www.isaserver.org/articles/2004domainnamesets.html
  • • if you suffer "creeping death" on an ISA server (more and more sites become unavailable, you may have inadvertantly blocked DNS queries going out from the SBS box
• • Group Policy Management
  • • user objects on SBS box are not in Users container, they are in MyBusiness->Users->SBSUsers
  • • If you want to add group policies to a set of users, you'll need to create a new Organizational Unit (OU) and add the GPO to the new OU
• • Windows Mobile 5 - blew right through this rather than getting actual demo, notes are published on a web site
• • Site to site scenario - running multiple servers in an SBS environment
  • • http://uksbsguy.com/r.ashx?7
• • Sharepoint goodies
  • • would have seen more on this if the SQL install/upgrade had completed before this part of the session, which was the end

Unfortunately, technical difficulties really hampered the effectiveness of this presentation.

On Amsterdam: Day 2 (wakeup through lunch)

Today we saw the demise of one of the sessions before anything started. Mark Mulvaney was supposed to do a talk on the Technology Assessment toolkit (which I was intending on attending) but he decided to shorten his presentation to about 10 minutes and include it in Susanne's presentation on the SBSpecialist community. So, that's where I ended up.

Susanne described the work she has done with the UK groups to get groups up and running and effective. To help explain the concept of community to those who aren't familiar with it, she enlisted the help of Tim Barrett and Christopher Rue, via prerecorded video capture, to help explain how they are getting value out of their community ties. Hopefully these videos can get posted somewhere, becuase not only are they good, but if you know Tim and Chris, you'll get even more enjoyment out of watching them (you'll just have to trust me on this).

Robbie Upshaw talked about the UK user groups from the MS perspective. It was interesting to hear his take on it, and I'd be interested to see what the US take on this is, too.

Mark Mulvaney did his bit on the technical assessment toolkit. Granted, he is right that a full session on the interface would have been dull and boring (yes, watching someone talking about cliking here and selecting this menu there would have been dull) but I would have been interested to hear more about the toolkit and practical applications. [Note: I've used the toolkit with one potential client and made the sale, so I have some familiarity with it, but I would have liked to hear more about it.]

I opted to skip the 70-282 training session (took the exam back in 2004) and was planning to attend Karl's network book discussion, but I could get absolutely no network connection in the room, and as I was checking e-mail right before heading in, I got an e-mail from one of my remote clients and needed to work with him.

So, had a nice lunch, catching up on e-mail and blogging, and am now getting ready for the afternoon sessions.

On Amsterdam: Day 1 (update from Marina)

Marina attended the Exchange session yesterday while I was off learning about SLAs. Here are her notes:

Exchange Sp2

Mobile: Local and remote wipe mobile device
Push email: ms mobile 5 with MSFP (Messaging Security Feature Pack)

Mailbox size: 75 gb

Spam: automatic updates imf
Sender ID filtering

Msfp locks security preventing to drop the sbs certificate: needs registry hack

Desktop: upgrade to activesync 4.1
Establish partnership with mobile
Use activesync 4.1 to install msfp
Choose certificate

Mobile administration web tool:
Creates another default website

Spf record puts ip address of mail server in dns. Receiving mailservers verify if IP is mailserver with Sender ID filtering.

Ta.

On Amsterdam: Day 1 (recap)

Wrapping up the summary of Day 1 right before I head off to sleep. Wireless access has been spotty in the conference area, so I haven't been able to update as much as I had planned.

Jeff Middleton's Swing Migration session followed a break following the keynote. If you haven't had a chance to experience Jeff's presentation on Swing, you are really missing something. I learn something new every time I sit through it.

Lunch was a disappointment. We met in the "vendor hall" and had sandwiches and drinks. I had something that I thing was sliced salmon with some sort of cream style spread on a bun. I'm a total novice to European cuisine, so I must admit a huge ignorance to what may or may not have been offered, but the supposed salmon looked more appealing than anything else that was offered, and that's not saying much. But again, that's probably more me than a reflection on the facilities.

After lunch, I sat through Mathew Dickerson's presentation on "Superior Service Level Agreements." As I'm still looking for solutions to this for my own business, I was curious as to his content. He related the history of how his company went through the transition of moving from a "break/fix" organization to a Service Level Agreement (aka Managed Services in other lingo). Honestly, not much new ground here, with one exception. The solution they developed (program structure, contracts, marketing materials, etc.) they have decided to license to other IT organizations. Hey, it beats reinventing the wheel. So, I'll be talking with him more about that in detail tomorrow.

A quick break and it was off to the MVP panel. There were 9 SBS MVPs in attendance at the conference today, and we got all of them in the front of the room. We had a slide showing that listed the MVPs and their countries of origin. Jeff gave a brief (that's funny to those of you who know Jeff) introduction to the discussion, then opened the room to the first question. This is the first question that was asked (and no, I'm not making this up - ask any of the MVPs who were in the room):

"I see Eriq Neale listed on the list. Is he here and which one is he?"

I can't recall the exact wording, because I wasn't expecting to BE the first question, but that was the jist of it. So I stood and acknowledged the asker, then commented on how that would be the easiest question we got to field all day. He actually asked a question about integrating Macs with Exchange, and I was able to ramble on about that for a while. The rest of the Q&A went smoothly, with all of the participants having an opportunity to field questions from the room. I'm not 100% certain that the European audience knew exactly what to do with an open forum for Q&A for the MVPs, because there were several long pauses waiting for questions. And even though we started the session late (and Harry gave us permission to go longer) we still ended up finishing at the originally scheduled time. Still, I think it was valuable for all who attended.

Following the MVP Q&A was the vendor forum where most of us mingled in the vendor room accompanied by finger food and alcoholic beverages. Then at 7, the activity broke. Some folks participated in the downtown Amsterdam scavenger hunt, but a few of us opted to remain behind. So I had a nice long dinner in the hotel restaurant with Jeff M, Mariette, Steve Foster, and Marina. Our dinner conversation spread to the bar after a while, and I left the proceedings around midnight local time.

Tomorrow I will probably attend the session on the Small Business Assessment toolkit, then the Network Documentation Workbook. Not sure what I'll attend in the afternoon. I'm sure it will depend on the mood and if there are any last minute schedule changes.

Until tomorrow...

On Amsterdam: Day 1 (wakeup through Keynote)

Wakeup call came in at 6:30am local time and, as confirmed by Susan B, I had to make sure to get some e-mail and blogging in before hitting the shower. Met up with Jeff M, Marina R, Steven T, and Monika or breakfast, then it was time to head upstairs and check in and get ready for the keynote.

No massive production like SMB Nation goes completely without some hiccups, and I was on the receiving end of one of them this morning - no printed name badge.

Got into the keynote room and found a seat near an outlet so Marina could hook up her power bar and get power for our laptops.

Harry started the keynote introducing the conference to the attendees. Since there are so many different countries represented by the attendees, Harry introduced the attendees by country (accompanied by the Olympic theme in the background). The UK had the largest contingent of attendees. Wow.

David Overton from MS UK delivered the keynote speech. Highlights include:

• • R2 is not necessarily a product that IT consultants are going to take to their existing customers for upgrade - it's really going to work best for new installs.
• • 6 new IDC whitepapers available - https://partner.microsoft.com/40025871 - for bundling technologies together in new sales/solutions for consultants to take to their customers
• • Technology Assessment Toolkit - not well heard of outside of US and UK; materials to help consultants present solutions to potential clients
• •SBS roadmap:
  1. 1. 1H 2005 - SBS 2003 SP1 released
  2. 2. 2H 2006 - SBS 2003 R2
  3. 3. Mid 2007 - 'Cougar' released - 64-bit ONLY
• • R2 contains server 2003, Exchange 2003 SP1, WSUS, ISA 2004 (premium only), Outlook 2003
• • big question relates to how the customer wants SQL - if they want SQL 2005, they need R2; otherwise, stick with SP1
• • WSUS is a key technology in R2 - an alternative way to get updates on servers and workstations
• • Exchange mailbox size increase included with Exchange SP2, along with new mobile enhancements (push e-mail, security, etc.)
• • Status reports now include information about system patches for updates - which systems are patches, which are not, and why (not on the network since X date, etc.)
• • you do not get Analytical Services with SQL 2005 workgroup in SBS 2003 R2
• • SQL 2005 WE in R2 does include Reporting Services, Report Builder, Management Studio, Log Shipping, Import/Export wizard, supports 3GB of ram (not limited to 2GB any more), native database encryption, SBS R2 CAL includes WG CAL
• • Expanded CAL rights
  • • access additional Exchange 2003 and SQL 2005 Workgroup in the R2 network (the servers must be joined to the SBS 2003 R2 network)
  • • customers will still need to purchase server licenses for Exchage and SQL, just not the CALs for them
  • • will not be required to purchase new CALs (ala SBS 2000 to SBS 2003)
• • how to upgrade to R2
  • • existing SBS customer if you have Software Assurance, you will not need to purchase a new server license, but your SA contract must be current when SBS 2003 R2 releases
  • • without SA, you will need to purchase a new license, can purchase the upgrade SKU
  • • new customers who purchase SBS 2003 between March 1 and 60 days after R2 releases, but you must purchase through OEM or system builder to get this (details to be confirmed)
• • can we stay on SQL 2000 after upgrading to R2? the answer is NO. (do not upgrade to R2 if you have to stay with SQL 2000)
• • what happens to server if I upgrade and I've already been running WSUS? rules are set to default, so back up your WSUS rules before upgrading because they will get overwritten

Harry introduced SMB Nation TV, a video blog site, smbnation.tv. Will start releasing a weekly show on the web, releasing April 15. In the first show includes an interview with Phillip Presnall (MS New Zealand), the Small Business Specialist manager in NZ. He discusses exercising discretion in SBSC enrollment and not just letting everyone into the SBSC "club." Also included is an interview with Debbie Engel from Portland, Oregon, Rodney Boyle and Henry Craven from Melbourne, Australia, Eric Ligman, from MS, Iam Phillips from London, Suzanne Lavine, Karl Palachuck, and Russell Haynes from San Fransisco, and finally Tim Loney and Beatrice Mulzer in Houston, TX. There was a lot of discussion among attendees related to the introductions Harry made to each of the video clips from the interviews that he presented.

On Amsterdam: Live Internet Radio

Just wrapped up the live broadcast of eOn Call, featuring an interview with Marina Roos of smallbizserver.net fame. This show almost didn't happen, though, becuase of internet connection issues. At 9am CDT, we connected via Gizmo to the braodcast booth in Denton and got a good signal going. Then the internet dropped out. It took 20 minutes and a panic call to the front desk, but eventually the network connection came back. Then about 15 minutes before the show, Gizmo freaked and while the folks in Denton could hear me, I couldn't hear them. Fortunately, that was resolved by hanging up and calling back.

The next connection stayed up all through the broadcast. And we didn't have any show host meltdowns, either. You'll be able to listen to the show at the eoncall.com web site here in a few days.

Now to grab dinner, chat a bit, and then get some much needed sleep. Conference officially starts tomorrow.

On Amsterdam: Leg 1

Only took me two tries to find a location in Chicago O'Haire airport that had a decent wireless signal. Fortunately I caught an earlier flight out of DFW, so I have longer to find a good wireless spot and catch up on e-mail and blogging.

Have I mentioned publicly how much I hate flying? I hate flying. Well, not flying per se, just takeoff, landing, and any time there's turbulence. Unfortunately on the flight out of DFW, there was turbulence during takeoff, some parts of the flight, and the last 10 minutes before landing. So I'm taking precautions before the second leg that takes me to Amsterdam.

I got to witness a few cool things on the flight in, however. We passed over one plane a few thousand feet below us going the opposite direction. if I hadn't looked out the window just then, I would have missed it. I also saw another flight running parallel to us, not headed to Chicago, and got a picture (I'll link the pic in when I get to Amsterdam and can buy a USB cable for my camera since I forgot to pack it). I also saw a number of brush fires (pics coming later) and realized that Chicago and its suburbs are definitely into baseball - I lost count of how many ballfields we flew over after about 55.

Next post from Amsterdam sometime on April 5. Probably before the radio broadcast (be sure to tune in to http://www.apostleradio.org at 10AM CDT) but that will depend on a lot of factors. Getting ready for the show is priority one tomorrow.

Until next time...