• Anonymous (unregistered) in reply to highphilosopher
    highphilosopher:
    Raid 1 is better than it used to be??? It hasn't changed. RAID 1 is mirrored.

    Maybe he was referring to raid systems that in the past didn't allowed parallel reads?

    Maybe you should read yours and some before you post.

  • Agitprop (unregistered) in reply to Burpy
    Burpy:
    - The haughty parasite spend trucks of company's money for nothing and fails. - The real computer guy takes all the risks, saves the day and shuts up. - The parasite is acclaimed

    Maybe one day people will understand that parasites can only survive if they find a resigned host to suck...

    Parasites? Someone has been playing a little too much Bioshock.

    vindico - What Frank Fontaine felt

  • (cs) in reply to dkf
    dkf:
    highphilosopher:
    As a lowly developer when I say, "we need these server specs to run this new app" management doesn't question it.
    Sure, but in these straitened times they still say “no” to the request for funds to get the hardware you say you need. There's what the app needs and there's what the app has got to be squeezed into because your capital budget was suddenly cut by 80%…
    If your company is too cheap to buy the hardware necessary to efficiently run your app, then you're probably not going to have a job much longer, as you're probably too expensive, too. Keep an eye out for contractors coming in to replace you.
    DWalker59:
    highphilosopher:
    I disagree. As a lowly developer when I say, "we need these server specs to run this new app" management doesn't question it. How many sysadmins do you know that have a server sitting at home because they "needed a new one for x project".

    No sysadmins that I know would lie, cheat, and steal like that. I'm serious -- I hate dishonesty. I don't have any servers at home because I claimed that we "needed a new one for x project". If you have done that, shame on you.

    Really??? Every sysadmin I've ever known has had a company provided server at home, purportedly to "test installations of new server packages in their free time". Most of them have hosted websites on them for a personal profit.

  • highphilosopher (unregistered) in reply to Anonymous
    Anonymous:
    highphilosopher:
    Raid 1 is better than it used to be??? It hasn't changed. RAID 1 is mirrored.

    Maybe he was referring to raid systems that in the past didn't allowed parallel reads?

    Maybe you should read yours and some before you post.

    Possibly he is, but lets look at this:

    1. RAID 1 refers to a conceptual way of arranging data on disks. It hasn't changed. It's a concept. If the concept changes, they call it something else.

    2. The RAID controllers have changed dramatically. If this was his intent, I would expect him to say something like: "RAID controllers have improved over previous designs" OR "RAID is now faster than it used to be"

    Instead he stated that a single concept had changed that hasn't. Basically that's like me saying that RAID 5 is faster than it was ten years ago. When I actually mean, you can now run a RAID 5 configuration on faster drives with a faster bus on a faster server. Which one would be the correct way to say it?

  • (cs) in reply to md5sum
    md5sum:
    If your company is too cheap to buy the hardware necessary to efficiently run your app, then you're probably not going to have a job much longer, as you're probably too expensive, too. Keep an eye out for contractors coming in to replace you.
    What I do isn't vulnerable that way as it happens, and budgeting (especially in large organizations) is more complex than you make out. There are a lot of places where salaries and capital budgets aren't interchangeable at all.
  • clive (unregistered) in reply to 3rd Ferguson
    3rd Ferguson:
    And if you are talking about physical data placement within a SAN you are about 10 physical layers away from the bits anyway and you'd almost always be better served by adding hardware that's closer to the database engine, meaning network, RAM and CPU cores.

    Databases get disk-bound very easily - network and CPU just won't help, and neither will RAM if what you're querying won't fit.

    1. Fix the queries
    2. Make the disk go faster

    in that order.

    Anybody in this discussion heard of WAFL?

  • Jo Woods (unregistered)

    No way dude, that is like totally insane!

    Lou www.privacy-news.us.tc

  • (cs) in reply to dkf
    dkf:
    ... budgeting (especially in large organizations) is more complex than you make out. There are a lot of places where salaries and capital budgets aren't interchangeable at all.
    No, but when the software you design suffers in performance due to management being too cheap to buy the required hardware, then suddenly your value to the company becomes less because your software contribution is reduced to a sub-par level. Thus, you become expendable, since "a product can be bought off-the-shelf that will run as good as your software".

    While the numbers may not be "interchangeable", they are directly proportional in almost every case due to the nature of our jobs.

    Hardware / 2 == Software / 2

    It's an inevitable decline.

  • Si (unregistered) in reply to dkf

    users are only happy when they're making stupid meaningless requests, and developers are only happy when they're running poorly performing development code on live servers. ;-) (I'm a DBA by the way)

  • Steve H (unregistered) in reply to BradC
    BradC:
    Ok, the DBA was a moran.

    Which Moran? Dylan? Caitlin? Surely not Kevin...

  • Matt (unregistered) in reply to J
    J:
    Who is the guy/gal approving these spending items, and how do I sucker my company into hiring that wasteful spending chump?

    Really, how the hell would all those drives get approval?

    Even assuming these were Velociraptors, order 10 drives like that is nothing compared to SANs and other enterprise hardware. So a few hard drives is nothing. And if they were standard consumer 7,200 RPM drives, then we're talking even LESS money.

  • Forumtroll (unregistered) in reply to Iie
    Iie:
    Let me guess, it was an ORACLE DBA, right?

    Yes it was. We had this fellow entertain (i.e. torment) us with his "ideas" about "optimalization". What the story fails to tell you, is that the same fellow found a way to taunt Oracle into throwing a critical error every time a package was loaded into memory (instead of the more usual "Package state has been discarded" error). This fellow did not stop at this point, but kept on increasing optimalization by enforcing the Oracle instance to unload a package from memory once the execution flow of the called procedure was finished. In case you wonder, this is where my company started experiencing an insane turnover.

    I have no idea how it is even humanly possible to make a living the way he did. I still cannot even grasp how he avoided all kinds of lawsuits due to this.

    I have no love for Oracle (actually I loathe it), but this certain fellow made my up-until-then distaste for Oracle seem benign compared to what I feel now. If someone encounters this DBA, the only viable option, apart from homicide, is to tender their resignation.

  • Steve H (unregistered)

    Does anyone have the magazine article scanned? Seems like that would settle it.

    Thought I'd mention that in case no one else does.

  • Paul (Another Paul, that is. ) (unregistered)

    Sounds like that other guy I can't stand... Whatshisface again? Burle-something? Says the same sort of stuff that once WAS valid, in 1836 or so...

    yeah, here's some: http://www.dba-oracle.com/t_raid.htm

  • Monkey (unregistered) in reply to BradC

    Real world testing trumps theory. Especially when the theory is wrong.

    People are inclined to believe what they read or believe in ideas instead of testing to see if it's really true.

    The certified DBA read somewhere that this crazy RAID setup would help (or he was just padding his consulting hours) (or the whole story could be made up) instead testing different configurations himself and seeing what worked best in the real world.

  • Anonymous (unregistered) in reply to DWalker59
    DWalker59:
    highphilosopher:
    I disagree. As a lowly developer when I say, "we need these server specs to run this new app" management doesn't question it. How many sysadmins do you know that have a server sitting at home because they "needed a new one for x project".

    No sysadmins that I know would lie, cheat, and steal like that. I'm serious -- I hate dishonesty. I don't have any servers at home because I claimed that we "needed a new one for x project". If you have done that, shame on you.

    Ahahaha, nice troll! I completely agree, sysadmins are very honest people and I certainly don't have a basement full of "repurposed" hardware at home!

    Meanwhile, back in the real world...

  • Paul (Another Paul) (unregistered) in reply to DWalker59

    If, as a sysadmin, you'd like to ship servers to your home, you should consider shipping a real life to your home instead.

  • Chris S (unregistered) in reply to BradC

    You have to be careful handing out unqualified advice like "RAID-5 = unhappy users". We use RAID5 for practically everything and with our light load we have happy users, happy DBAs, and a happy SA.

  • sy (unregistered) in reply to jimicus

    In our SANs we've got huge write caches. Our writes are magnitudes faster than our reads. I'd assume in most decent raid configs you've got a similar situation.

  • Anonymous (unregistered) in reply to Paul (Another Paul)
    Paul (Another Paul):
    If, as a sysadmin, you'd like to ship servers to your home, you should consider shipping a real life to your home instead.
    You college kids will learn to love stolen hardware eventually. It's merely a matter of time.
  • Matt (unregistered)

    The concept of the outer tracks of the disk giving better throughput than the inner tracks is perfectly valid - Pillar Data do exactly this on their storage hardware as part of the QoS system in the SAN. Even a quick and dirty test with a 7.2krpm SATA drive shows >100MB/s on the outside tracks and closer to 50MB/s on the inner tracks.

    Of course once you start reading and writing from all over the disk at random, of course you'll see a penalty, but the idea of only using the outer tracks of 10krpm disks and adding more spindles should result in some pretty astounding performance.

  • RadarBob (unregistered)

    The first moronity, foreshadowing all the subsequent moronity, was declaring that "the outer edge of the disk is fastest." Well, linear velocity, yes; but it's the same rotational velocity. I/O is the same rate regardless of which cylinder is being read.

  • Robert (unregistered) in reply to Anon

    Totally agree. The biggest moron in the story is Paul.

  • (cs) in reply to Matthew
    Matthew:
    OK, someone please find the original article in the magazine :) Must read.

    It might help to search by author. Start with the name "Mikkel."

    Mikkel:
    This smells fishy, Paul claims he could achieve orders of magnitude more performance by putting up a mirrored harddrive compared to many disks with dedicated partitions, and it was an I/O bottleneck? Even if the data was spanned and the DBA used some very old way of doing stuff, the fact is there are more drives available and should have had higher I/O throughput.

    Also, Paul should be fired on the spot for doing this, he is messing around with something he clearly doesn't understand (not that the DBA was any wiser, it is however the DBAs responsability), having tools intentionally report faulty information will make debugging extremely problematic.

  • (cs) in reply to Another DBA comment.
    Another DBA comment.:
    Ok, the 'certified' DBA was going way overboard, I'll give it that. 60 partitions? Unless we're dealing with a multi-terabyte database, that would seriously scare me.

    However, there is a LOT of research into disk layout for database servers. Please look into it before you criticize either the DBA or Paul, too much here. There are valid reasons for placing logfiles on separate partitions (which I do by default), and breaking out large tables or indexes into separate files and disks.

    If you really want to know the details, google "Paul Randal SQL Server" and read up on it. (He's the guy who wrote DBCC CHECKDB for Microsoft, and I think he knows a thing or two about on-disk structures for SQL Server...)

    HTH

    Stop confusing partitions with drives. Separating workloads across drives improves performance. Separating workloads across partitions reduces performance by encouraging thrashing. The biggest error I see in database IO subsystem tuning is taking a recommendation that uses the word "drive" and substituting the word "partition". This seems to be what happened in today's WTF.

    I also don't understand why people don't benchmark these configurations before making assertions.

  • Ugh (unregistered) in reply to BradC

    *moron

    Yes the disk configuration is relevant but your statement lacks any understanding of the disk subsystem. I could put 100 LUNs on a single RAID-10 array and you would have horrible performance. Some times I wish some DBAs could be SAN admins for a couple months. :(

  • A.Nonymous (unregistered)

    Imbecility on this scale should not be anonymous: if that guys managed to infect some "serious" publication with its "knowledge", his name should be made public.

  • (cs) in reply to Matt
    Matt:
    The concept of the outer tracks of the disk giving better throughput than the inner tracks is perfectly valid - Pillar Data do exactly this on their storage hardware as part of the QoS system in the SAN. Even a quick and dirty test with a 7.2krpm SATA drive shows >100MB/s on the outside tracks and closer to 50MB/s on the inner tracks.

    True, but it is only the sequential read/write speed that get increased, not the seek time, and I think that the db would be limited by seek time in most cases.

    And I still find the thought of doing raid on partitions interesting, but I can't imagine how the dba thought it would be faster then doing it on the entire disk.

  • (cs) in reply to RadarBob
    RadarBob:
    The first moronity, foreshadowing all the subsequent moronity, was declaring that "the outer edge of the disk is fastest." Well, linear velocity, yes; but it's the same rotational velocity. I/O is the same rate regardless of which cylinder is being read.
    The problem isn't I/O... it's seek... Read: http://partition.radified.com/partitioning_2.htm
  • Brandon (unregistered) in reply to BradC
    BradC:
    Ok, the DBA was a moran. But that doesn't mean that ALL DBAs who claim to know something about physical partitioning are idiots.

    Even with a SAN, the physical configuration is certainly relevant to throughput.

    RAID-10 on dedicated LUNs = happy users, happy DBAs RAID-5 on shared LUNs = unhappy users, unhappy DBAs

    so was he a mormon or a moron? I can't tell if you are being a jerk or just condescending

  • Agitprop (unregistered) in reply to DWalker59
    DWalker59:
    No sysadmins that I know would lie, cheat, and steal like that. I'm serious -- I hate dishonesty. I don't have any servers at home because I claimed that we "needed a new one for x project". If you have done that, shame on you.

    I can't say that I've ever arranged to have a developer tell management that the specs on the server weren't great enough, but then our database is puny and all of our applications run on the clients.

    However I have inherited servers from work, especially as we virtualized. Of course this was back in the heady days of being a tenant and having my utilities included in my rent. Now I'm down to 2.

  • alister (unregistered)

    As a system administrator I'd say using Hardware RAID will accelerate the heck out of the system far more than any partitioning system. Plus having separate arrays (and battery backed cache controllers) for the OS/SWAP/Logs/Databases (if you really can push the boat out get separate array for tables). Additionally if you can ensure that you have enough PCI(e/X) lines so that you don't have a bus bandwidth issue that's even better. Built one box a few years ago, three separate PCI-X buses, two cache controller and the network card on separate buses system ran sweet, front screen access was slow though.

  • SR (unregistered) in reply to Steve H
    Steve H:
    BradC:
    Ok, the DBA was a moran.

    Which Moran? Dylan? Caitlin? Surely not Kevin...

    I don't think that sarcasm was warranted but massive respect for getting Keving Moran into a TDWTF comment! :o)

  • Anonymously Yours (unregistered) in reply to highphilosopher
    highphilosopher:
    Anonymously Yours:
    Anonymous:
    I'm rooting for Paul in this story but I have to ask - why didn't he just show his greatly improved two-drive solution to the boss?

    "Hey boss, you know all these problems we've been having? Turns out the DBA doesn't know how to set up an efficient drive array, I've done it myself with just two drives and the performance is spot on. Also, you can get a refund for all these unused 10,000rpm drives he made you buy. So anyway, about that raise...".

    Then the DBA accuses Paul of sabotaging the previous system while beating his chest about his extensive experience. The DBA had to have some clout if he managed to:

    1. convince management to buy a crapload of 7200 RPM disks
    2. convince management to buy a 10000 RPM disks to replace those
    3. convince management to buy even yet still more 10000 RPM disks when all of his previous purchases failed, and
    4. then skipped off on vacation while the system he was solely responsible for was bogged down

    I disagree. As a lowly developer when I say, "we need these server specs to run this new app" management doesn't question it. How many sysadmins do you know that have a server sitting at home because they "needed a new one for x project".

    Those are both good points (I'm not even a sys admin and I am guilty of having an unused server at home). Still, can you honestly say Mr. "Fifteen Years Optimization Experience" isn't going to throw Paul under the bus in a heartbeat? Paul doesn't seem particularly good at handling minor confrontations. I can't picture him recovering from being knifed to the back repeatedly.

  • NewbiusMaximus (unregistered) in reply to Anon
    Anon:
    So TRWTF is that Paul allowed this total nonsense article to go out in a trade magazine... This highlights the real lack of scholarship in computer science circles.
    Please don't conflate an unnamed DBA trade magazine with computer science scholarship. I know CS academics sometimes make bad programmers and sysadmins, but there's no way an article about this kind of bullshit would pass the laugh test at any reputable CS journal.
  • Anonymous (unregistered)

    This sounds like some of the articles that I've seen in MSDN and CodeProject. I can't count the number of times I've had to rip out code copied from one of these "expert" provided articles. Any idiot can write an article or a tech book apparently.

  • (cs) in reply to Ugh
    Ugh:
    *moron

    Yes the disk configuration is relevant but your statement lacks any understanding of the disk subsystem. I could put 100 LUNs on a single RAID-10 array and you would have horrible performance. Some times I wish some DBAs could be SAN admins for a couple months. :(

    Obviously my RAID 5 vs 10 statement was simplistic, but you're supporting my underlying point: just because its a SAN doesn't mean its automatically "good". You have to have a SAN administrator that does more than simply allocate the 500GB of space you request--they have to do so with appropriate considerations for the performance requirements of the server in question (something which does not happen in my workplace).

    And for anyone who didn't get it, "moran" wasn't a typo, it was a reference to the internet meme:

    [image]
  • Sigivald (unregistered) in reply to Anon

    So TRWTF is that Paul allowed this total nonsense article to go out in a trade magazine, read by other DBAs, perpetuating a myth about partitioning.

    How was Paul supposed to stop it? The article doesn't even claim Paul knew about it in advance of publication.

    Yeah, blame him for not knowing the incompetent DBA was going to publish something and somehow stopping him (how?).

    Good call, there.

  • J (unregistered) in reply to Jo Woods
    Jo Woods:
    No way dude, that is like totally insane!

    Lou www.privacy-news.us.tc

    Did a bot just get past the capcha software? This definitely looks like advertising bot spam.

  • Some Moran (unregistered) in reply to BradC

    I think you mean he was a morOn...

  • Anon (unregistered) in reply to NewbiusMaximus
    NewbiusMaximus:
    Anon:
    So TRWTF is that Paul allowed this total nonsense article to go out in a trade magazine... This highlights the real lack of scholarship in computer science circles.
    Please don't conflate an unnamed DBA trade magazine with computer science scholarship. I know CS academics sometimes make bad programmers and sysadmins, but there's no way an article about this kind of bullshit would pass the laugh test at any reputable CS journal.

    Sorry, but I've seen stuff submitted to CS journals and been to CS conferences, and most of the bullshit I see there wouldn't pass the laugh test at even a low end hard science journal (physics, chemistry, biology, etc).

  • lol (unregistered) in reply to BradC

    A moran? The irony here is thrilling

  • JohnsonRodpecus (unregistered)

    The purpose of this article was what???

  • Antonio Tejada (unregistered) in reply to RadarBob
    RadarBob:
    The first moronity, foreshadowing all the subsequent moronity, was declaring that "the outer edge of the disk is fastest." Well, linear velocity, yes; but it's the same rotational velocity. I/O is the same rate regardless of which cylinder is being read.
    Uh, when's the last time you did any disk cloning? Ever? There's a huge difference in the sustained throughput of the beginning of the disk (the edge) and the end (the inner track) — it can be almost 50% slower on a 3.5" disk.
  • Garp (unregistered) in reply to Anonymous

    Not sure what SANs you've been using, but the ones I've come into contact with actively work to optimize data storage based on performance profiling, moving data where appropriate to ensure that you get as good performance as possible.

  • Mike (unregistered) in reply to BradC

    All the partitioning BS is moronic, but I always thought that using only the outer 25% actually does help boost performance, known as "short-stroking."

  • Todd C (unregistered)

    Am I the only one to question the 'Certifiable DBA's' claim that the "outer ring" was the only decent usable space on the drive because it is furthest distance from the center?

    True, a point on the outside of the disc travels at a greater linear speed than a point nearer the center, but both points travel at the same angular velocity 7,200 or 10K rpm.

    Not only does this DBA know more (or so he thinks) than the sysadmin, he also knows things about platten hardware manufacturing than Maxtor and Seagate.

    Hey Mr. DBA, when you were 12 and listened to vinyl records, (remember those things with one loooooong spiral groove?) did you only listen to the first song because the others were considered of 'impure quality'?

  • (cs) in reply to lol
    Some Moran:
    I think you mean he was a morOn...
    lol:
    A moran? The irony here is thrilling

    Have none of you seen that photo floating around the internet?

    It was even linked to already, just scroll up a bit.

  • PG (unregistered)

    Started as a sysadmin, many many moons ago, then started doing DBA work with Oracle 5.1, then backup in SAN and sysadmin work.

    RAID-5 bad for Oracle, not true. At one time Oracle published that, but later removed that statement.

    What RAID-5 is bad for is small sequential writes to a file that you never go back and read.... Hey that sounds like just like an on-line redo log! To make matters worse is when someone takes a large number of physical drives and puts them in one RAID-5 set. All the extra reads to then compute parity will kill you.

    It's not just Oracle, I had someone put an Exchange server and it's database on a SAN I controlled. He claimed he need at least 14 drives in the RAID set to get good performance. The array by default only used 4 drives in a RAID-5 set because of how the back end end data channels worked. The thing was many times faster than any of his other setups where he setup a RAID array himself.

    The biggest issue I have seen and still have to deal with my local DBAs is the the difference between a mount point they see and physical spindles. The Oracle doc just doesn't make this clear. Now add in all the other layers of volume managers, SANs, LUNs, RAID sets, and it gets confusing for them fast.

    Not saying that all DBAs are the problem, a SAN guy recently didn't understand the issue either and he would create a BIG RAID set, then break that in into a few LUNs. Each of these LUNs went to a different database on different servers. So heavy IO on server A hurt server B. With the DBAs and customers screaming, one week that he was on vacation we moved a few LUNs around and fixed the problem.

    He still doesn't understand why what hid did caused a problem.

    It real simple

    (usable space) * (performance) * (easy admin) / $$$$ = a constant.

  • zeno (unregistered)

    Lol @ those commenting about linear/rotation velocity. In say, 1/4 of a rotation, where will the head cover more surface? Inside? Outside? Doesn't matter? The only way reading/writing would have same speed on the inner rings as the outer rings would be if packing density on the rings varied. While this would make the reading/writing speed constant, it would also be a waste of disk area if the head can support reading more at those speeds (7.2, 10k).

Leave a comment on “The Certified DBA”

Log In or post as a guest

Replying to comment #:

« Return to Article