is happily being the wheel rather than a rusty old spoke
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Register for free to join our community of investors and share your ideas. You will also get access to streaming quotes, interactive charts, trades, portfolio, live options flow and more tools.
Is Top Gear completely off the air?
I've pre-ordered the DVD at Amazon, but no release date has been announced.
Have you seen Talladega Nights yet, Tom? I'm thinking us road course folks are going to get a lot more chuckles out of it than your average bear. Everyone I know has been anxiously awaiting it. I hope to be seeing it with my family tomorrow night.
If you've seen it already, don't spoil the surprise by telling me which direction the turns are in.
Until I saw the part about that he was "heading for turn 6", I'd envisioned the deer simply standing on the track *in* 6, which would definitely do the trick as that one's blind and there's certainly not enough room to avoid a deer (or car) that suddenly enters your line of sight as you crest the hill.
I'm guessing it must've crossed just past the bridge, otherwise he likely would've seen it as he exited 5.
Sounds like they've got a good handle on him at the hospital, and it's probably fortunate that he likely won't remember what happened. That kind of thing would make it awfully tough to get back on the horse.
101 here today. That's not so bad, but the heat index is between 110 and 115.
And I'm out mowing and working on the lake. Drinking and sweating gallons.
Time to check on what I've got the db server doing (while cooling off a bit), then get back out to it.
We're unable to have the message search functions be realtime right now because the database server can't currently handle the extra load along with the greater load it's seeing from the huge increase in traffic recently.
We've ordered a new database server and the new version of SQL Server, which, among other things, has vastly improved full-text search functionality and performance.
In the meantime, we're updating the full-text indexes afterhours and working on other ways to reduce the database's load so it can return to realtime full-text indexing, and if that can't be accomplished soon, I'll try to move that function to SI's database server, which has tons of excess capacity going unused.
Wow! Truly a scary thing to imagine happening at our speeds in fully-enclosed cars, let alone so exposed in a Champ Car!
Nah. Somehow I put real bullets in the gun you were playing with. Now I need to figure out what I did.
As I've just written you privately, and am mentioning publicly for everyone else's benefit, we think what you're reporting and the occasional outage of Favorites are directly related. Your message likely being the cause. Well, our web-server's handling of the message when you try to submit it.
If people have been responding to your messages, they'll still be in your MailBox if you haven't read them. Those don't get cleared out until you actually read the message.
Would've been the same answer. I'm really puzzled by this one. The exact same routine is not working in the development environment now. Trying to get a consensus in favor of rebooting the webserver.
An essay entitled "Beats the Hell Outta Me!!!"
Favorites.asp was broke. I fixed it. Do I know what was wrong? No. I made seemingly useless changes and it started working.
The end.
FWIW, today was the first time I'd edited Favorites.asp in ages.
He's got it turned on or he wouldn't have seen the posts missing in the Board list of messages.
I strongly suspect it's a nuance of Ignore I never did address.
The "Clear" link is what you want.
I can't remember now why we chose "Clear" rather than the more obvious "Reset".
When I click on FAVORITES and I see one of the forums has 170 NEW POSTS, I then click on the 170 count and I read that single post. Then if I immediately go back to FAVORITES the NEW POSTS count is now set to zero. It should show 169 since I read only that one post.
That's not a bug. It's always worked that way.
There are three ways to reset a board's new post count to zero:
1. Hit the link showing the number of new messages.
2. Read the last post in a board.
3. Hit the "Reset" link for that board on Favorites or hit the "Clear" link at the top of that column.
The reason we don't set the unread counter to change when you read a post is because you might end up going to a post on a Favorites board via a link in another message.
If you hit the new post count to get to a board and read just one post and want to leave the rest marked unread, just hit the "Mark as Last Read" link on that post.
Also, there is still a problem with IGNORED posters. On another forum, I clicked on the NEW POSTS count and it took me directly to a post by a user I have IGNORED. I went to his profile and it says UNIGNORE THIS POSTER so I interpret that as confirmation that he is on my Ignore list. I then go to the forum's list of posts and that specific post is not shown, as it should be NOT shown. So posts by Ignored posters are still being shown when accessed off of the NEW POSTS link from FAVORITES.
I'll look into this one. It's entirely possible that bug has always been present, if it happens to be the first unread post and you access the thread via the new post count on Favorites. Because handling ignores is a very difficult and expensive (in computing resources) process, and it may be a subtlety I never nailed down.
Next time that happens, hit "Next" on the post, then hit "Previous" and see if it takes you to the same post. If it doesn't, that would narrow it down to a problem (perhaps that's always been the case) only with the new posts link. In which case, it'll have to wait, as I've got far larger fish to fry at the moment.
But if you hit "Next", then "Previous" and are taken to a post you shouldn't be seeing, that's a big enough fish that it needs more immediate frying.
No. First, it's not "low horsepower". We've got a ton of it. But not enough for it to do everything we'd like it to do.
The ignore function should be working now pretty much the same way it did a year ago, warts and all. If memory serves, there was still some fine-tuning needed.
Looking fixed now? I see what I did. I got within a mile of an ignore function.
Okay, I see what you're saying.
read_msg.asp seems oblivious to the Ignore On/Off toggle and always acts like it's on.
This is gonna be interesting, because I really didn't think I'd gone within a mile of anything ignore-related in my recent changes.
test again
That post is hosed. Even says so now.
Wrote it from the development directory mid-development.
On the other problem you guys were seeing nearly universally, one of the changes I'm making right now has to do with how signatures are dealt with. You were getting the error message because I kinda spaced out that lots of people don't have signatures and wasn't accounting for that correctly.
Better now?
hosed post
test
Is this thing on?
test
ztest
And it'll make my life a lot easier, too. The two things iHub most needs right now in a db server is plenty of horsepower and SQL Server 2K5.
Current box is SQL 2000, 2 Xeon 2.8's, 4GB memory, and about 70gig of hard drive.
A pair of 3.0Ghz 5160's, SQL 2K5, 8GB memory (they're pricing me 16 to see if I want to up the ante before the machine gets built) and 700GB of what's supposed to be much faster hard drive. And everything will be 64-bit.
I'm guessing I'm looking at something along the lines of 2-4 times as much horsepower, which will definitely help a lot, and SQL 2K5 will be wonderful because of how much better and more efficiently it does Full-Text Indexing. Not only is it supposed to be able to update indexes "orders of magnitude faster", it gives good visibility into the catalogs so I can, for example, print out the 1000 most frequently-indexed words and see which ones are unnecessary, resulting in smaller catalogs, faster updating, and faster searching.
And the main problem with iHub right now is that with the sudden increase in traffic, we've got enough horsepower to deal with the traffic or maintain realtime full-text search, but not both. Though I'm still working on ways to do both before the new box arrives since it likely won't be installed and put into production until early September.
We already have a new webserver (pair of 3.8Ghz Xeons) but have yet to put it into production. We've had a number of false starts with it, so we're holding off now until the new db arrives and will put them both into production at the same time.
The inhouse phrase we use is "If Churak's happy, the site's happy." <g>
We need to hire someone in Greenland so normal morning hours for them will be wee hours for us.
But then it'd be so much more difficult to kill them if they call at 3:00 AM with a non-emergency.
I doubt I'll be getting up very early in the morning. If you or Matt sees any sign of delays approaching unacceptable in the morning, please give me a call.
I don't wanna give away ALL of the business side of things.
Darn. Okay, lemme go have a look.
Edit: Did I say "Darn"? I meant "Yay!!!"
You just single-handedly answered a question that really had Dave and I scratching our heads. I think.
That's the field I was referring to where I couldn't stand that it was a big text field with possible values of "Yes" and "No" when it should've been a bit field with 0 or 1. I made the change to the table and to the only routines I could find that could modify it.
But we were getting emailed error notifications that "Yes" and/or "No" were trying to get stuffed into a bit field.
Turns out chgquote.asp was a place that could happen.
Let 'em. <g> And maybe they'll have pity and subscribe like crazy to help us fund it.
A BUNCH!!!
FYI, I just fired off another Public Message catalog incremental population and have it set to try to maintain a realtime catalog. We'll see how it does tomorrow morning.
BTW, here are the specs on the new server:
Base Unit:
Dual Core Xeon Processor 5160 4MB Cache, 3.00GHz, 1333MHz FSB, PE 2950 (222-3390)
Processor:
Dual Core Xeon 2nd Processor 5160, 4MB Cache, 3.00GHz 1333MHz FSB, PE 2950 (465-5552)
Memory:
8GB 533MHz (4x2GB), Dual Ranked DIMMs (311-5731)
Keyboard:
No Keyboard Selected (310-5017)
Monitor:
NO MONITOR OPTION (320-2968)
Video Card:
Broadcom Dual Port TCP/IP Enabled, Microsoft OS Only (430-1497)
Video Memory:
Riser with 3 PCIe Slots for PowerEdge 2950 (320-4607)
Hard Drive:
146GB, SAS, 3.5-inch 15K RPM Hard Drive (341-3031)
Hard Drive Controller:
PERC 5/i, x6 Backplane Integrated Controller Card (341-3066)
Floppy Disk Drive:
No Floppy Drive for x6 Backplane (341-3685)
Operating System:
Windows Server 2003 R2 Standard x64 Edition,Includes 5 CALs (420-5798)
Mouse:
Mouse Option None (310-0024)
NIC:
Dual Embedded Broadcom NetXtreme II 5708 Gigabit Ethernet NIC (430-1764)
Modem:
Dell Remote Access Card, 5th Generation for PowerEdge Remote Management (313-3923)
CD-ROM or DVD-ROM Drive:
24X IDE CD-RW/DVD ROM Drive for PowerEdge 2950 (313-3934)
Sound Card:
No Bezel Option (313-0869)
Speakers:
1x6 Backplane for 3.5-inch Hard Drives (311-5747)
Documentation Diskette:
Electronic Documentation and OpenManage CD Kit, PE2950 (310-7415)
Additional Storage Products:
146GB, SAS, 3.5-inch 15K RPM Hard Drive (341-3031)
Feature
Integrated SAS/SATA RAID 5 PERC 5/i Integrated (341-3060)
Feature
No Rack Rails Included (310-7411)
Service:
Dell Hardware Warranty Plus Onsite Service Inital YR (984-1399)
Service:
Basic Enterprise Support: Business Hrs 5X10 Next Business Day Onsite Service Post Problem Diagnosis Init YR (970-4070)
Service:
Dell Hardware Warranty, Extended YR (984-1417)
Service:
Basic Enterprise Support: Business Hrs 5X10 Next Business Day Onsite Service Post Problem Diagnosis 2YR Ext (960-8162)
Service:
BASIC Enterprise Support: Business Hrs 5x10 Hardware Only Tech Phone Support, 3Yr, Declined Software Support (960-8192)
Installation:
On-Site Installation Declined (900-9997)
Misc:
Redundant Power Supply with Y-Cord for PowerEdge 2950 (310-7421)
146GB, SAS, 3.5-inch 15K RPM Hard Drive (341-3031)
146GB, SAS, 3.5-inch 15K RPM Hard Drive (341-3031)
146GB, SAS, 3.5-inch 15K RPM Hard Drive (341-3031)
146GB, SAS, 3.5-inch 15K RPM Hard Drive (341-3031)
If I'm not mistaken, the 5160's got 4MB of shared cache and the 5060 has 2 separate 2MB caches, right?
All I really based my selection on was them telling me that despite the clock speed (which I seriously balked at), I'd get much better performance out of the Dual Core at that speed than the fastest non-DC Xeon I could get.
Was I told correctly? I usually trust Dell on these things, although they did incorrectly tell me some time back that I could just increase the size of my RAID5 volume by adding hard drives to it on the fly. Nope. Adding drives means a complete reformat/reinstall of everything.
How's a pair of 5160's at 3.0Ghz sound?
Speaking as the guy whose main job is to ensure that my, Matt's, and Dave's passionate hobby is also a viable business, I can categorically state that we don't consider free members to be "freeloaders".
Not unless they make a lot of work for us. :)
I should point out, though, that I've done considerable analysis regarding free vs paid membership and the following things are true:
1. Our ad revenues are just over half of our total income.
2. Free members outnumber paying members by about 3 to 1.
3. From solely the perspective of the bank account, the averaging paying member contributes about 35 cents per day to the coffers. The average free member (through ad views), about 4 cents. That number really is a little tough to pin down, though. That's just based on a free member's pageviews. When a free member writes a lot of good posts and they get read a lot by other free members, more ad revenue. And more people buying subscriptions so they can more easily read a lot of posts.
Bottom line is that ponying up for a subscription, contributing content for others to read, and/or seeing and actually paying attention to ads and clicking ones that look like they're worthwhile, it's all money in the bank, and it's all good.
That problem has cleared now, right? I think it should be working now and you likely caught it right in the middle of me changing the structure of a related table. Found some old stuff and couldn't bear to leave it as-is, even though it's not costly. It's just plain WRONG. A field storing the text "Yes" or "No" when it should've been storing a boolean (0=false, 1=true).
Everyone, what's the consensus on how the site did today?
The histogram looked pretty nice until about half an hour before the close and is just now starting to taper off. Even with that spike, it looked very tolerable, but I'd still like to find some more headroom before re-enabling full-text indexing.
I probably published about a dozen new versions of read_msg throughout the day. It doesn't appear I've caused any outages, and hopefully I didn't cause any problems to crop up in scenarios I didn't test.
I've gone through every line of code up to line 220 so far (out of 339 currently -- was about 400 this morning) in read_msg and held the magnifying glass up to each and every one that has any interaction with the database. Most of the inefficiencies I've found were quite small, but with how frequently the webserver sends queries to the db, it does tend to add up.
In some cases I've revamped routines to run a tiny bit faster and in the most recent change, I completely did away with a separate, very tiny and inexpensive (but remember -- called a hundred or so times per second) query that simply wasn't needed because I already had the info I needed from another query. And I think I'm going to be able to zap another one shortly by adding just one field to the main query.