Slashdot Mirror
Worlds Slowest NT Server
Thom writes "NerdPerfect is holding a contest to find the NT server with the slowest reboot time. The best (worst?) time so far is 49 minutes, 13 seconds. Go check it out."
← Back to Stories (view on slashdot.org)
← Back to Stories (view on slashdot.org)
Our exchange server has taken well over an hour to shut down before. And, of course, since it's an NT box with a lot of traffic hitting it, it's gotta be rebooted every couple of months.
Now that they're moving everything to Active Directory services, just imagine, you can have your website's registered users, your email users AND your in-house people all mucking up the same "technology" at the same time. And the API is almost useable! Lucky us, I tell ya.
What I see - and I see it very often - is that a NT machine will start up pretty fast, slow down over a the course of a few hours of hard work and finally all but stop. At that point, the system may not have technically bluescreened, but it sure isn't any fun to work on. The only cure that I know of is a reboot.
I never see this problem on Unix systems - I think it has to do with the way Windows applications interact with the OS, with all the shared gunk that clogs up everything.
D
----
...something an NT server will be good at! ;)
"Total destruction the only solution" - Bob Marley
This sounds like a cool contest, but how are the times being checked? Or is this like the Darwin awards, where if you can come up with a reasonable answer, you're good to go?
Not that I don't believe that it could take Windows NT 49 minutes to reboot, I'd just like to see it.
by Mike Buddha -- Someday the mountain might get him, but the law never will.
... exhibitionist masochist network administrators. Is this news, or daytime television?
I have a copy of NT 3.51, running on a 12.2mhz OC'd V-series x86 clone, with 4M of memory and dual 400M IBM full-height MFM/RLL drives.
At thirty+ minutes to restart with no services/processes running, I think it could win. Anyone ideas on how to make it slower? (I'm not going to engage in anything like the Exchange kill timout bug. Thats cheap.) Can't strip out any of the HW, I'm already below the minimum req's..
.sig: Now legally binding!
I'm sure I could load NT on my 486 currently running Linux. It'll take at least 5 minutes to boot right out of the box. If I add a few service packs and essential services, it should get up to at least 20... What else?
æeee!
It's an interesting, and almost plausible theory, but Microsoft themselves recommends AGAINST this.
;-), the Microsoft recommendation was to never have more than either 2 or 4 thousand users/domain. (and this was the official microsoft documentation). If you ever check out the network structure of a large company (someplace like Merrill Lynch where Private Client Services has something like 60,000 nodes on the network) you'll note that it is NOT a single domain.
When I got my MCSE (know thy enemy
I can't see ANY reason why you wouldn't be better off to have a PDC and BDC at each satellite and then set up trusted domains.
If you have a single domain then that means that the central administrator would have WAY too much access. 'the admin walked away from his desk and forgot to log out of the god account, so I ran a program which changed 10% of all the passwords, and randomly deleted 2 files off of every machine'. This would be a Bad Thing(tm).
(And now, I'll see massive holes ripped in this postulate. Bring it on!)
"So, what do you want to hack for, Bickle?"
"I can't sleep nights."
"They got porno theaters for that."
"Whatever happened to fair use?"
-- Duff-Man
Two years ago I was the Admin of a NT box with Exchange 5.5 on it. It had about 50 mailboxes on it, but all the mail was kept on the server (In one large file, no less). It would routinely take about 30-40 minutes to reboot, depending on what exchange was up to at the current moment.
Unfortunately, we had a power failure overnight and the server went down. Hard. Very Hard. Exchange was NOT happy about having open files all over the place. Exchange would take forever to try and bring up all its messaging services, and usually fail in thr process of initializing the internet gateway. Before we blew the whole mess away, it was taking almost 2 hours to reboot.
Surprisingly, the box was a PPro 200 with 224MB memory (128+64+32)
"I may disagree with what you have to say, but I will defend to the death your right to say it"
"I disapprove of what you say, but I will defend to the death your right to say it."
- Evelyn Beatrice Hall
linux could _easily_ reach boot times that windows NT could never dream of, because you have the source.
Just open up the kernel source, and somewhere in the code for the opening crawl insert code that simply runs an empty while() loop until GetDateTime() is equal to January 7, 2004.
Hell.. long as you're at it, you might as well make it a for(;;);. Infinite boot time! woohoo!
Irritable, left-wing and possibly humorous bumper stickers and t-shirts
No, not for my NT server to reboot. :) :) oh, and I have a Mac Classic in the bedroom as a cheap notepad/xterm for my shell account, and it boots in twelve seconds (not much to boot). ;) "So what? They're only customers."
That's how long it took for my old powermac to quit Netscape, shut down PPP, reboot, start up PPP, launch Netscape, dial in, load Slashdot, load this page and return to where I was reading
For some reason I find this whole discussion insanely amusing
Steve Jobs once threw a fit at Apple developers, getting them to make the Macintosh boot faster. "Twenty seconds??? There are X million customers! If you waste a single second you're wasting time equal to X many entire lives! Are you ready to waste that many lives to your laziness???" (not a quote, just paraphrase). Maybe we ought to send him to Redmond to yell at some people. I guess the Microsofties don't mind the toll of lives they waste
The t-shirt probably reads:
"I am the NT Admin whose machine booted the slowest in the big contest. Hire me, I am an expert"
It's to be worn at job interviews. And when applying for food stamps.
All you have to do is set the timeout on NTLoader to be ridiculously high...like 65535 seconds.
Let's face it: this is nothing more than an attempt by an unknown website to get a little free media coverage, and they've obviously been successful.
Great contest idea, though a T-Shirt for a prize seems like a pretty half-assed prize imho :) Here's how I would handle the prizes if were running things:
Me: Congratulations, you have won the contest with an amazingly pathetic reboot time of 4 Hours, 57 Min! (or something to that extent)
Contestant: WOOHOO... What did I win? What did I win?
Me: Oh we have a special prize for you... this brand new, top of the line server... running on your favorite OS... WINDOWS NT!!!!!!!!!!!
Contestant: NOOOOOOOOOOOOOO!!!!!!! AGGGHHHHHHHH!!!!!!!!!! (goes insane and flys into a fit of violent spasms before falling to the ground, dead)
Now wouldn't that be funny and ironic?
Take two linux computers, put 10 filesystems on each. Now NFS mount the 10 fs from one machine on the other, and vice versa. How do you shut them both down, without logging on as root and unmounting all the filesystems manually first, because the power has failed perhaps?
In case you don't know, if you have some NFS filesystems mounted, and try to reboot when the NFS server is down, it will take about 5 minutes per filesystem before umount -a times out.
Okay, explain to me how this makes Windows NT a bad OS ? (Apart from the obvious). Of course if you jack up and load the fucker with lots of apps, it'll take forever to reboot, the same with linux and all other os's.
/etc/hosts name.
Hell, do you even realize how long it takes to fsck and parallelize drives upward of 20gig on a linux box ???. And if the name server is down, sendmail takes forever until it times out to take the
Like I said, stupid, retarded, and it only serves as to satisfy all those rabid NT-haters out there.
One could easily have the login authenticated accross an artifically slow link creating very long login times. I think they should take this out of the contest.
Then again, this isn't scientific, and is good clean FUD, so sure, whatever!
-AP
Not so, my man. This is only measuring one thing: a contest with slow NT reboot times.
:)
If you want a contest for slow server boot times, so be it. Put up a webpage and ask for submissions.
My linux box is running quite a few services, and it isn't a monster like some of the hardware for some of those NT boxes, and it boots up quite nicely, thank you.
fsck can take forever. Fortunately it doesn't always run. However, NT checking partitions takes a while too. Sendmail doesn't take *that* long to timeout IMO, but I could be wrong here.
A 50 minute boot time sounds like Exchange is a bad *app* to me, though. I wonder what it does.
What do you think creates rabid NT-haters? NT! It's DOS, VMS, and Windows, all rolled into one! How monolithic and anti-UNIX can you get? It doesn't tell you anything, it's slow, it doesn't make decent-sized partitions, it has weird driver support, it doesn't support a lot of windows extensions from the past 4 years, it's insecure, it's bloated, and it's generally annoying. Therefore, the occasional anti-NT page is entertainment.
If you can write a good, sensible anti-UNIX or anti-whatever page, and get it posted, go right ahead. Start with X, it annoys me sometimes. Even if it does have all the features Windows Terminal Server wishes it has.
---
pb Reply rather than vaguely moderate me.
pb Reply or e-mail; don't vaguely moderate.
but 49 min to reboot would make me have to hurt someone.
I would imagine that this isn't a workstation. It could have a ton of different servers loading in that time: HTTP, FTP, SQL, etc. Maybe it has a lot of drive space.
There are many reasons why an NT server would be that slow.
I'm not saying that's acceptable, however. My main Linux server reboots in, I guess, around <90 seconds. That's with a lot of daemons, too.
There is a known problem with Microsoft Exchange 5.0 (and 5.5) that in certain circumstances can cause an NT machine to take a LONG time to reboot (approx. 1-2 hours). I believe that a machine with Exchange will be the winner because of this problem (which I spent days with MS trying to track down this problem).
This problem is apparently due to a service deadlock (i.e. the service does not respond to the kill command). There is a registry entry (what a shock!) that dictates how long NT will wait for a service to shutdown. Of course, this is a PER service command. If you have a deadlocked service subsystem (as can occur in certain circumstances with Exchange), it will take approximately * to bring the system down. I believe the default is 5 minutes. Changing it to ten or twenty seconds does a lot of good. =)
Looking at the website, the leader with 49 min has Exchange running on it. Maybe we should notify him of this registry entry... Nah...
Later,
Justin
Mu. P.S. The address you see is real. =)
I think that's the slowest processor that will run NT, and I just happen to have one sitting at work. I'll just pop that into the Compaq 486 w/24Megs of RAM, and we'll see how long it takes to INSTALL, let alone go through reboots... :-)
Thank God it's Friday, and I can actually slack off and do this all day!
Exchange Server 5.5 can take a *very* long time to boot (ours takes about 5 mins end-to-end, 250 users) especially if you haven't purged down and defragmented the information stores.
:)
49 minutes indicates an Exchange box that isn't getting the care and attention it requires.
As for the 18-minuter; well, that's probably one unresponsive box. Remember, Microsoft recommend one server per server application; now, please buy Small Business Server, which runs SQL, Proxy, Transaction, Exchange Servers all one one box
--
Peter
...Its no problem to make a server running any OS rebooting really slow if you try.
Now, the contest ought to be for the slowest rebooting time on a machine set up by some certified MS whatever, in a place where the downtime really hurts. Like in a cold storage. "Oh sorry, you will just have to leave the 10 tons of butter out here in the sun. Computer problems you know..." (true story)
All opinions are my own - until criticized
Just another reason I'll stick to Solaris, thankyouverymuch. UFS logging being a wonderful thing, especially when you have a 60GB RAID-5 volume.
--
DHCP addrsssing on a multihomed server? Did this guy read the instructions or help files? Can he read. MS says not to do that with their clients.
Yeah, I ran into that. It's really too bad because the NT DNS was really nice in that it registered the DHCP client's names via WINS.
So, Instead I ended up moving all the services to a multihomed Linux box runing ISC DHCPd, BIND, and Samba with a perl script that would suck the entries from Samba/DHCP and put them in the dns.cache file.
The NT box boots alot faster now, and the Linux box does not seem to care...
What's the point of this? To show that NT takes a long time to boot up? To make NT look bad?
It doesn't do either. The organisers come across as being incredibly immature, and the existence of such a lame contest is hardly a credit to the Linux community.
...is a really lame excuse. Isn't one of the primary tentant that Windows is supposed to be easy to use? I have seen admins with credentials up their arm, screw something up and have to reinstall from scratch. Us "windows developers" pratically have to do it every day.
In theory, one could solve all your problems by knowing how every thing ticks, but most of time we work with a lack of information.
I'll tell you what - you release Windows in source code form and I will document it the way it should be - completely. No more obscure registry switches, no more half brain head services. I would love to know why SCM does the brain head things it does.
He probably didn't shut the exchange services down in the correct order before shutting down exchange.
:-)
I have a pair of scripts on all three of our exchange servers, and use one of them to shut them down first before rebooting the server (the other one is for restarting the services, but is rarely used).
Here's the shutdown script:
@echo off
rem batch file to stop Microsoft exchange
echo Stopping services
net stop MSExchangeIMC
net stop MSExchangeMTA
net stop MSExchangeES
net stop MSExchangeIS
net stop MSExchangeDS
net stop MSExchangeSA
echo system down
As for longest shutdown, I did have once where one person here decided to reboot the server at the end of the day (and didn't know about the scripts), and it was still trying to stop the services at 8am (15 hours later!).
-- trust me, I'm a Viking
Java gaming nut - http://www.retep.org/ or for the rail http://uktra.in/
There was a bug in the Spooler service. It wasn't apparent until Exchange 5.x was installed on the system. The spooler would not shutdown because it was waiting for a connection that never came. Installing Exchange modifies the global shutdown WaitToKillService timeout to ~15 minutes. This was done to allow Exchange time to shutdown its databases without fear of corruption for being killed. If you install SP4 or greater, this is fixed. Long Reboots is no more funny than PCWeek not installing the security patches for Red Hat and calling it a fair security test. The only good thing that should be taken away from this is that all admins need to stay on top of the latest patches for their platform.
Ever tried running the system info tools? You can get graphs of everything from vm swaps to cpu util in user vs protected mode to packet loss to ... well you get the picture. That app is the one thing I really miss about NT...
it doesn't make decent-sized partitions
Huh?? NTFS supports 16exabyte partitions. You have a device bigger than that??
it has weird driver support
Weird in that there are drivers shipped in the box with virtually any hardware device you can buy (except of course some video capture products and parallel port devices which I agree is a bit crap) and you don't have to write them yourself? :-)
it doesn't support a lot of windows extensions from the past 4 years
It doesn't support DirectX over version 3. Period. And Win2K will fix that...but hey only Linux people are allowed to use the phrase "that's in the next release" aren't they?
it's insecure, it's bloated, and it's generally annoying.
And I can't be bothered any more...
Seriously I'm not having a dig at the linux/solaris/be/*bsd/amigaOS/whateverOS guys & gals out there. Just don't fight FUD with FUD...it only lowers yourself to their level. And you really don't want that. NT's not too bad if you know you're way around it, and are aware of it's limitations. Different OS's are good for different things. Nothing is perfect....chill :-)
---- Den ene knappen er powerknapp, den andre er Bender voice knapp "Bite My Shiny Metal Ass"
You'd be surprised.. I have a system that fsck's 6 8.5 gig drives in parallel in under 60 seconds. (I think it is about that timeframe, it has been 65 days of uptime since it last booted.) Running a single fsck, on one drive (95% full) takes 35 seconds real time (as just measured)
:)
But lets put it this way, that system takes less time to boot up with an fsck than my personal machine which has one 90% full 5gb partitian formatted with the defaults among its drives.
Those 6 drives are intended for big datafiles, so I make the blocksize 4kb as compared to the default 1kb. This means 1/4 the number if indirect blocks. Also, since its big files, there aren't a lot of directories so it doesn't have to scan through them. Those directories that there are only require 1/4 the time to read because they are 1/4 the number of blocks. (I assume that sequential reads are relatively free because seek time should be the main cost.)
Then, I formatted it at one inode per 128kb, as compared to the default of one inode for every 4kb. This drops the number of inodes fsck has to read off the disk by a factor of 30 to only 69632 (compared to >2 million in the default format). (FSCK *has* to read off each inode, to check if the file is orphaned.) I think that this here is the main speed improvement. It also frees an extra 120mb of drivespace. (drivespace that an fsck would have to read in.)
Remember, since you can never create new inodes, so the default format always starts you with a large amount of them, usually an insane number of extra ones.
What I'm saying is to format your drives appropriately for their planned usage. FSCK's don't always have to be that painful.
If you need that huge quantity of extra inodes and also have to save the data, than you are screwed. If you don't need all the inodes, then formatting a filesystem without them is a lot more effecient. And if you don't need to save the data (say, a cache server), then it may be faster to just reformat it upon each boot.
If none of the above apply, then yes, with the current version of ext2, it is very painful to fsck a 20gb filesystem that has 3 million files on it.
These are very good tips. (In particular, I had missed the "1 inode per 4K bytes" bit in my brief encounter with the mke2fs manpage, but indeed that's the default.)
BTW, for the curious, the magic incantation that produces the above set of parameters is like so: mke2fs -b 4096 -i 131072 /dev/ whatever
I have a 17Gb hard-drive that I use for CD-ROM images for when I build music-compilation CDRs. These tips would work great on this HD, particularly for the odd case when I have to fsck. Note that a 4Kb cluster size should also speed up most other operations on large files, particularly those which require synchronous metadata updates, such as rm's.
(Note: I've only had one Linux kernel crash that I can remember since getting this drive, due to some strange interaction between burning a CD, playing a WAV file and insmod'ing the floppy driver from kerneld on a kernel that would have preferred kmod. The rest have been stupids like bumping the power switch, having a UPS go bad, etc.)
--Joe--
Program Intellivision!
-----------
"You can't shake the Devil's hand and say you're only kidding."
REM ha.. I win!
pause
There's a sound reason for this, it's so you can have central administration for all users regardless of physical location. This is useful if you have resources spread throughout your organisation which must be shared to a wide range of different users, or if you have a central HR organisation. It means a single sign on and easy ACL administration.
The correct way to do it is to have a PDC+BDC in your main office, and a BDC in each remote office. You want another BDC in each office for every 2000 users. This arrangement scales up to about 40,000 users in a single domain, then you'll need to move to a multiple master domain.
...it's been going on for many years. The fastest boot sequence of any reasonably general purpose OS I've encountered was on Univac 494/Omega that I worked on back in the early 70s (when it was already nearing the end of its lifetime): given suitable peripherals it could boot and have the online workload running again in under 10 seconds - because it had been designed for a target market which could not tolerate extended outages. And because the machine had a maximum of 262Kwords of memory, and ran at only about 1.2Mips, there was a premium on careful programming that used resources efficiently. Even then, though, it was clear that the trend was towards the cost of implementation and maintainance of application software exceeding that of the raw processing power, so that it became increasingly cost-effective to use less expert implementors and to provide them with tools that traded implementation time against run-time performance. And that is a perfectly reasonable and defensible way of proceeding, provided that the various costs are taken fully into account.
:-(
However, there's a big difference between making a considered guestimate of the price of processing power over the likely lifecyle of a project and using these resources accordingly, and making the mistake of thinking that "cheap" resources are "free", and placing no constraints on their use (or, equivalently, blindly assuming that available systems will be fast enough to run the product by the time it is released). In the old days, if you did this you were liable to fail to meet contractual performance criteria, so there was a fairly direct feedback mechanism that kept product performance reasonably compatible with the capabilities of the platforms on which the products ran. More recently, with commodity pricing for software products, this feedback loop has been broken, and the costs of poor implementation practices are now born disproportionally by the customers - whether by being forced to replace systems more frequently, or with outages after major software failures becoming longer, or by finding that the newest products increasingly manifest inconsistent and unpredictable failure modes.
It's interesting, in this regard, that there have been rumours that MS may be considering leasing of its products rather than (or more likely in addition to) the customary "shrink-wrapped" sales. Major corporate users might just be interested in this approach, if it gave them some contractual leverage when the fitness for use of a particular product was less than satisfactory.
Just don't hold your breath
What is funny about this type of thing is that IE5 actually INCREASES the reboot time on all machines, becouse i automagically loads all of the IE DLL's into memory. Then they proceed to do ads stating how fast IE is.. Well sure, they preinit everything at startup, and your time is moved from when you click to when you turn on the machine. Not to mention the simply swap issue under low memory conditions..
-- I'm the root of all that's evil, but you can call me cookie..
I have heard of a machine that took six DAYS to boot. Running a netscape server (I think a news server product?). Lots of files (millions?) on a very big, very slow drive. took six days to finish running chkdsk when it restarted.