1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Is your company using false economy like mine?

Discussion in 'The Lounge - Off Topic' started by greenbrucelee, Jun 7, 2010.

  1. greenbrucelee
    Highly Decorated Member Award

    greenbrucelee Zettabyte Poster

    14,283
    254
    329
    My company is always going for the cheapest of options to save money but it always back fires and ends up costing them more.

    I have just discovered that the six yearl old database servers use a raid 5 setup when really they should be mirrored and now they are faced with having to go mirrored and therefore costing them more money than it would of if they had implented a mirrored setup in the first place.

    does your company do stuff like this?
     
    Certifications: A+, N+, MCDST, Security+, 70-270
    WIP: 70-620 or 70-680?
  2. zebulebu

    zebulebu Terabyte Poster

    3,748
    330
    187
    Mirrored where? RAID5 is the most common RAID level for the data in databases. RAID1 is only really common for the O/S. What you tend to find in most places is that they put the data on RAID5 to start with, then later, when it starts getting mullered, they move to RAID10 or similar to increase performance. I'd be surprised if they're moving the data to RAID1 - they're more likely moving it to RAID10 but didn't do so in the beginning because it's hideously expensive. If it's the O/S that wasn't mirrored, then they're idiots.
     
    Certifications: A few
    WIP: None - f*** 'em
  3. dales

    dales Gigabyte Poster

    1,997
    46
    97
    Yep cheapest option always wins and us techs actually have to shoehorn it into some kind of working order. I suspect most companies buy things for the lowest figures. When you have sales reps saying that their products can even make tea and biscuits for you mgmt rarely know any different and go by the pound signs.
     
    Certifications: vExpert 2014+2015+2016,VCP-DT,CCE-V, CCE-AD, CCP-AD, CCEE, CCAA XenApp, CCA Netscaler, XenApp 6.5, XenDesktop 5 & Xenserver 6,VCP3+5,VTSP,MCSA MCDST MCP A+ ITIL F
    WIP: Nothing
  4. greenbrucelee
    Highly Decorated Member Award

    greenbrucelee Zettabyte Poster

    14,283
    254
    329
    sorry I meant it was the OS side that isn't mirrored and now its screwed up and its gonna cost them ****loads to get sorted.
     
    Certifications: A+, N+, MCDST, Security+, 70-270
    WIP: 70-620 or 70-680?
  5. greenbrucelee
    Highly Decorated Member Award

    greenbrucelee Zettabyte Poster

    14,283
    254
    329
    I think what the problem is in my place is the IT director used to be an accountant, the IT manager doesn't give a **** and everyone else has to pick up the pieces.
     
    Certifications: A+, N+, MCDST, Security+, 70-270
    WIP: 70-620 or 70-680?
  6. Phoenix
    Honorary Member

    Phoenix 53656e696f7220 4d6f64

    5,726
    175
    221
    generally that mentality is two fold

    it's a business goal to cut operating costs and increase the bottom line
    and if you don't have techs skilled enough at actually getting the BUSINESS benefits of making one choice over the other to the top brass, then the top brass will obviously go with the cheapest sustainable option

    the sad fact is, 99% of IT professionals are a) not professional and b) have absolutely no idea of business, they just think its all about 1s and 0s


    I can't imagine why a busted OS is going to cost so much to fix unless the DB vendor charges consultancy to do a bunch of gay registry hacks to make there hack of a database work better

    i could wipe all my database OS's and have the databases online before the wipe began using log mirroring and some fancy DNS, rebuilding an OS should NOT be hard work


    as for the Raid debate
    Zeb is correct
    RAID 5 is the most common place for databases
    RAID 1 for the log files
    RAID 1 for the OS

    with technology such as aggregates from NetApp this becomes even less important, its all about the data
    hell exchange 2010 can be installed on JBOD, as long as your using DAGs to duplicate the DATA
     
    Certifications: MCSE, MCITP, VCP
    WIP: > 0
  7. greenbrucelee
    Highly Decorated Member Award

    greenbrucelee Zettabyte Poster

    14,283
    254
    329
    point taken but it's got to a point because it hasn't been maintained properly and the fact that the OS hasn't been mirrored that the whole thing needs a complete rebuild.

    I understand that a business is supposed to make money and attempt to save money but when it gets to a point when the initial point of saving money ends up costing more in the long run which is stupid.
     
    Certifications: A+, N+, MCDST, Security+, 70-270
    WIP: 70-620 or 70-680?
  8. Phoenix
    Honorary Member

    Phoenix 53656e696f7220 4d6f64

    5,726
    175
    221
    but that may still be more beneficial to the business from a PnL and cash flow perspective
    I've got numerous clients that can't afford to do all my recommendations on day 1, but can continue in the next budget year even at increased cost
    from their perspective this is better for their business, it keeps the cash outlay lower even though its drawn out

    its no different than me buying stuff on a credit card
    i know i pay more, but i base my purchase decision off the total cost (after interest) and spread the payments out, this keeps my cash flow sound without hitting me too hard for a large purchase

    one could say saving works too for that, but im impatient and will happily pay a little more to get it now ;)
     
    Certifications: MCSE, MCITP, VCP
    WIP: > 0
  9. Phoenix
    Honorary Member

    Phoenix 53656e696f7220 4d6f64

    5,726
    175
    221
    and seriously
    not doing RAID1 on an OS is NOT generally a business move, its a tech **** up
    no way any business will deny the purchase of 1 additional hard disk for OS redundancy (were not talking moving RAID5 to RAID10 here, that could be pricey!)

    but i can point out plenty of techs that would overlook that specific build step ;)
     
    Certifications: MCSE, MCITP, VCP
    WIP: > 0
  10. craigie

    craigie Terabyte Poster

    3,020
    173
    155
    We commonly use the following:

    RAID 1 - OS
    RAID 5 - Data

    If the clients a cheap arse then one RAID 5.
     
    Certifications: CCA | CCENT | CCNA | CCNA:S | HP APC | HP ASE | ITILv3 | MCP | MCDST | MCITP: EA | MCTS:Vista | MCTS:Exch '07 | MCSA 2003 | MCSA:M 2003 | MCSA 2008 | MCSE | VCP5-DT | VCP4-DCV | VCP5-DCV | VCAP5-DCA | VCAP5-DCD | VMTSP | VTSP 4 | VTSP 5
  11. zebulebu

    zebulebu Terabyte Poster

    3,748
    330
    187
    Yeah, I agree. I don't think it would have been a budget issue in the first instance - more the tech in charge of the build either didn't give a toss about what they were doing, or didn't know what they were doing.
     
    Certifications: A few
    WIP: None - f*** 'em
  12. Theprof

    Theprof Petabyte Poster Forum Leader

    4,570
    68
    196
    I'll have to be honest, us too we look at the price and try to get the cheapest stuff, it's just how IT from a business prespective works. You have to understand that IT, in the majority of the companies is an expense and I'll though we are not bringing in the money, we do have to support the company enough for them to make money and that's something that a lot of techs don't understand. Which is why when we're researching new products we have to write proposals and explain how it's going to benefit us from a Total Cost of Ownership point of view and time management, etc...

    Just to give you an example. We're in the midst of upgrading all the workstations in the comany to Windows 7 and having a 400 user company it can get pretty price when it comes to hardware and software so we arranged a deal with upper management to buy 10 workstations every month of a year, so we don't drop a lot of money in one shot.

    It's not fun sometimes, especially when something breaks due to cheap products but we're forced to work with what we've got!
     
    Certifications: A+ | CCA | CCAA | Network+ | MCDST | MCSA | MCP (270, 271, 272, 290, 291) | MCTS (70-662, 70-663) | MCITP:EMA | VCA-DCV/Cloud/WM | VTSP | VCP5-DT | VCP5-DCV
    WIP: VCAP5-DCA/DCD | EMCCA
  13. Gingerdave

    Gingerdave Megabyte Poster

    991
    44
    74
    Umm me in that case, AFAIK none of the servers in the 50 plus I look after have a Raid1 setup for the OS :unsure

    Is this really that bad a move?
     
    Certifications: A+,MCP, MCDST, VCP5 /VCP-DV 5, MCTS AD+ Net Inf 2008, MCSA 2008
    WIP: MCSA 2012
  14. gosh1976

    gosh1976 Kilobyte Poster

    337
    18
    35
    I would imagine cost benefit analysis skills are a skill set lacking in many IT departments! Hell probably in many companies in general!
     
    Certifications: A+, Net+, MCDST, CCENT, MCTS: Win 7 Configuring, CCNA
  15. Josiahb

    Josiahb Gigabyte Poster

    1,336
    39
    97
    Thank christ... I read the first post of this thread and thought I'd got it horrendously wrong somewhere......

    I've got into the habit of basically not telling my boss what the cheapest option is, I tell him what the most expensive is and then really impress him by lopping a lump off the price while buying exactly what I wanted to in the first place. This very rarely fails.

    Got to agree though that your current problem doesn't sound like a budgetary issue, I would bet there's a whole host of techs out there who have rolled out servers with this set up. It makes perfect sense until the whole thing goes to hell in a handbasket.
     
    Certifications: A+, Network+, MCDST, ACA – Mac Integration 10.10
  16. Phoenix
    Honorary Member

    Phoenix 53656e696f7220 4d6f64

    5,726
    175
    221
    It's not an issue i come across much anymore, 90% of the servers i play with are virtual :)
     
    Certifications: MCSE, MCITP, VCP
    WIP: > 0
  17. zebulebu

    zebulebu Terabyte Poster

    3,748
    330
    187
    Depends. If you have enough disks in the server, you should always run the O/S on a seperate mirrored pair. I've seen people suffer from punctured stripes before - if this happens with everything on the same array (O/S, Data, Apps etc) you lose the lot, whereas if you run a separate O/S mirror and data array the chances of BOTH O/S disks and two disks in a RAID 5 array punking at the same time are remote to say the least (and indicate problems with the controller, rather than the disks)

    If you only have a limited number of disks, of course, then you've not got much choice. I've seen SBS servers running on 3 disk RAID5 arrays (which perform brutally) when there are seven slots free in the server chassis. That's just stupid. If you've only got a pizza box with three/four disk slots, then you haven't got a choice - though painful experience has taught me that underspeccing servers for clients leads to contract cancellations or non-renewals because the services running on them are perceived to be ****e.
     
    Certifications: A few
    WIP: None - f*** 'em
  18. greenbrucelee
    Highly Decorated Member Award

    greenbrucelee Zettabyte Poster

    14,283
    254
    329
    well this is sort of how it is there are plenty of disks and plenty of free slots 7 or 8 and the whole thing is gonna come crashing down if something isn't sorted soon. The performance is bad to say the least.
     
    Certifications: A+, N+, MCDST, Security+, 70-270
    WIP: 70-620 or 70-680?
  19. Gingerdave

    Gingerdave Megabyte Poster

    991
    44
    74

    Thanks Zeb.

    all the servers here which are critial generally store the data on external arrays or have a full on live backup server so if it all keels over and dies I just kick the users onto the other server while I rebuild it.

    Not perfect but that is the model I have inherited.

    Lots to think about now. Thanks to all.
     
    Certifications: A+,MCP, MCDST, VCP5 /VCP-DV 5, MCTS AD+ Net Inf 2008, MCSA 2008
    WIP: MCSA 2012

Share This Page

Loading...