Okay, so Windows Server 2003 SP1 was released Wednesday night. I started my deployment Thursday night, and spent ~8 hours getting my 3 servers installed and working perfectly with it.
Notes on the upgrade:
Well, it actually fixed bugs I had seen! Specifically, the following KB support articles detail the issues resolved:
827991
833734
883944
824905
829082
833708
837117
890477
It's great to see issues that I've encountered actually fixed.
Now, for the issues I ran into. First, using my Exchange server as the first server to install on and try and troubleshoot issues on was a bad idea. Since it takes around a half hour to reboot, the diagnosis of problems, and then the confirmation that the fixes actually worked took A LONG time. About 2/3 of the time I needed for the install was waiting for the actual install process to run, and then rebooting the servers.
I ran into 2 DCOM issues. First, the Exchange Information Store needed to have the permissions on component {9DA0E106-86CE-11D1-8699-00C04FB98036} modified. For some reason, after the upgrade it had no permissions set. Adding SYSTEM with "Local Launch" and "Local Activation" resolved that problem. There was also a DCOM error with a product that I'm beta testing, but that had probably been occurring before the service pack install. My guess is I just didn't notice it until I was trying to figure out why DCOM was generating errors in the event log for Exchange.
The biggest problem I ran into was a result of following the advice/templates from the "Windows Server 2003 Security Guide" from MS. Specifically, one of the recommendations is to use security policy to enforce permissions on services. Well, the original configuration breaks the Windows Time Service, and generated a lot of spurious event log entries. The exact error being generated was about the "local service account failed to logon". It turns out the permissions on the Net Logon service need to be changed from (System:Full; Administrators:Full; INTERACTIVE:Read) to (System:Full; Administrators:Full; INTERACTIVE:Read; NETWORK SERVICE:Read; LOCAL SERVICE:Read). After that change, all the problems pretty much disappeared. Don't know what specifically changed in the service pack to require that change...it just seems kind of obscure.
Hmm...I guess there really were only those two issues. The netlogon one was a bitch to find though (what do you mean it can't login!!! It already IS!!!!! Look at the TCP/IP NetBIOS Helper!).
For those of you that don't follow the blogsphere closely, here's a few links to other blog postings I found interesting yesterday (and that others might find even remotely interesting).
Adventures in product testing: Candles that catch fire
Incidentally, for those of you that remember the Windows 95 Power Toys and Kernel Toys, well, Raymond Chen (author of that blog) wrote those. He's also one of the reasons I don't think I could ever get a job at MS...the people there are just waaaaay too smart.
La Vida Robot
(Chris Sells is also notable for entirely different reasons, but mostly for just being "a character". Or something. Just subscribe to both blogs. If you're looking to subscribe to blogs that is.)
My commentary has now ruined what was meant to be a short, pointful post.
Monday evening, while doing a last minute look through in OneNote of my CS 251 notes, I noticed that my HP TC1000's pen was behaving a bit sluggishly. I shrugged it off, since there were MUCH more important things at hand.
So I went to ECON 251 the next morning and the pen worked great for 7 seconds or so, then just stopped responding. I tried unscrewing it and making sure the battery and everything was seated properly, then used it successfully for ~20 seconds. This made drawing graphs fun. I tried repeating the procedure, and got the pen to work for really short bursts like that, then just gave up after 15 minutes and went to keyboard only mode.
Now, having a Tablet PC and having to use it ONLY with a keyboard sucks (no mouse...just the keyboard...there's been a history of problems when trying to use the HP keyboard for the last year and a half).
So I scurried back to the dorm, got in my car, and made a trip to "Jeffrey's (un)Authorized HP Parts Warehouse and Service Depot" (aka "home", "parent's house", <address removed to protect the guilty/innocent/undefined>). After much exciting driving (for those not in Indiana, that is an example of "sarcasm"), the 6 miles of road were traversed. And then the real challenge, finding the replacement part amongst the shelf after shelf of other parts (ie, junk), began.
<much adventure skipped because I file stuff away well and found the pen & batteries immediately, but feel free to make up whatever heroic narrative you want and insert it here>
So with the replacement parts in hand, I return to the dorm, insert the new AAAA battery, and the old pen comes right to life. And the Tablet PC is restored to full functionality!
<disclaimer>There may be language considered offensive contained within. If you are apt to being offended by that kind of stuff, please stop reading now. I'm also writing this ~20 minutes after finishing the exam, so it may be a bit incoherent.</disclaimer>
OMFG! I can't believe he did that to us on the exam...now my head just hurts. It sucks.
Not like last year's exam at all...about 50% more coding, making more like 2 years ago. Also making it longer. And more difficult. Anyway, I didn't study the AVL Tree code stuff that much...I mean, it's 300+ lines, so there's no way he's going to expect us to write all that on the exam, right? Oops...questions 14, 15....1/3 of the points on the exam...bleh! That one screwed me over...
Hashes, heaps, nice simple short data structures to implement. Great for testing, so the thinking went... Well, it was half right, we had to implement a hash table...
And I was right, he did put Red-Black Tree deletion on the exam, after telling us in class that it wouldn't be on it. On the bright side, he told us "not to do #10" during the exam, but still...it was there.
I guess the AVL Tree stuff didn't go that badly. I wrote out the cases and figured them out. It took a while though. We started early, and he let us finish late, so 2:40 for the exam. bleh. Damn you Gustavo!
(yay, now I feel better)
Muhahahahahah....
My evil-half-thought-out-plan of not adding anyone to my Face Book Friends thingy unless they discover me first (via whatever means) is progressing well...in a week I've accumulated 4 friends!
Now, I'm beginning to think this is a stupid plan, but for the sake of I'm too lazy to change it right now because I have lots of other stuff to do, I'm not going to change it.
In any case, so far people have found out about it via this blog and word-of-mouth (I'm guessing). So let's start a second phase of this experiment: if you have IE (ya ya, I KNOW I need to get comments fixed for non-IE browsers, I just don't know how to fix it), reply with a comment to this post (click view comments at the bottom of the default postview, which takes you to the individual post view, where you click the big "Post comment" button) saying how you found out about my Face Book entry. I'm also going to add on the rule that only people who have added me to their friends should reply (although I probably wont filter the responses anyway).
Between me and the internet sits my firewall, MS ISA 2004 SP1 (currently). One of ISA's features is integrated caching (it can also do prefetching and scheduled updates, but I find those features to be more of a pain to manage then they're worth).
Now, caching is a great technology and all, and it can really improve performance, but in the past it's been kind of annoying. Emil (no web site :() runs www.boilerinvasion.org, and so I occasionally try and troubleshoot stuff for him. Except ISA has this annoying habit of caching that site like crazy. Consequently, I'm usually more of a hindrance than a benefit when it comes to debugging/opinionating stuff.
Fortunately, getting BoilerInvasion to work most of the time was fairly easy — just set the site to tell intermediaries to not cache it (I'm not quite sure what performance impact that's had...any comments Emil?).
However, this morning I ran into a wierd bug, which is probably due to caching. When visiting www.weather.gov to get the temperature before going to class, I refreshed the page (it had been left open from when I went to my earlier class). And the infromation didn't change. So I figured it was just IE being wierd and forced a refresh (CTRL+F5), bypassing the IE cache. And then the results were updated, displaying a nice 52° F. Then I hit CTRL+F5 again. And then the results changed back to the earlier values. Wow! It dropped 12° in the space of a refresh!
So I sat there for a minute or so just force refreshing the page, and it kept alternating back and forth.
Cache can be annoying sometimes. (of course now it's working fine).
Yay! It's warm! And the weather is nice!
And I'm spending the day inside cramming for a CS251 exam. Go CS. Or something. On the bright side, looking at the previous exams greatly increased my confidence. But the pressure is still there.
I just finished up Spring Break last week, and it more or less sucked. I was sick for half of it. Right before, there was a big CS project due that I completely bombed. Yay...now I get to take C S 251 again! (at least probably...although that Improbabilty Drive that should be appearing any yesterday now may/may not change that).
I also got a good bit of rest and relaxation, which is always good. And I got to see a FIRST ( www.usfirst.org, www.boilerinvasion.org) Regional competition. Kind of sad to see the lack of actual practical applicability from my high school (let's just say they didn't do so well in the competition). Better luck next year 461!
The Intranet project at work finally shipped last Thursday before last (or something...2005-03-10). Yay!
Interestingly enough, the usage stats show my load generating app with the largest amount of hits (by an order of magnitude), followed by...me. I'll see how things go over time.
Anyway, I'm now actively pursueing Method 1, including acquisition of the probability manipulation device. I mean, the movie is coming out soon, so there must have been a prop, and it's pretty improbable that a prop would actually be a working Improbability Drive system...and it'd be even more improbable if it just happened to pop into my hands yesterday all of a sudden...
So now all I have to do is sit back and wait. Muhahahahaha...
|
About the author

Jeffrey Stults is a software developer currently in Portland, Oregon. He is contactable at:
stultsj@ntldr.net
Archive
| | Sun | Mon | Tue | Wed | Thu | Fri | Sat |
|---|
| 27 | 28 | 29 | 30 | 31 | 1 | 2 | | 3 | 4 | 5 | 6 | 7 | 8 | 9 | | 10 | 11 | 12 | 13 | 14 | 15 | 16 | | 17 | 18 | 19 | 20 | 21 | 22 | 23 | | 24 | 25 | 26 | 27 | 28 | 29 | 30 | | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
Disclaimer
Disclaimer
The opinions expressed herein are my own personal opinions and do not represent
my employer's view in any way.
© Copyright
2012
Jeffrey Stults, Jr.
Statistics
Total Posts: 256 This Year: 0 This Month: 0 This Week: 0 Comments: 23
Utilities
Pick a theme:
Sign In
|