wigginsd88

Distinguished
Jul 9, 2010
19
0
18,510
Dear Community,

It has come to my attention that the Dells in my office building, Tetra Tech, have all been plagued with a common problem: Hard Drive failure. Whether or not this was a general fault in the HDD design, or overall, the degree of craftsmenship Dell puts into their Upper-End OEM builds has been weighed and balanced by my office manager, an intelligent man, but helpless when it comes to computers. He suggested terminating the contract we have with them, but I convinced him that before this was done, it may be worth looking into the componets themselves.

The fact that it is always seeming to be the Power Supply Units, or the HDDs failing suggests to me that it may be either to much of a load for either or both. We regularly transmit in the upwards of 1272 GBs in any given 9 hour day, plus our CAD department tends to leave their workstations on for overnight plotting (printing large documents, maps, etc). If anyone has heard of this trend, please speak up. I'd rather not blindy trash Dell let alone terminate a contract 3.5 years early on a "Hunch". If it helps, the HDDs are 7200 Rpm Sata drives running in Spanning Raid (Seagate Barracudas) ande PSU's are Antec EA-500D Earthwatts (post OEM by the IT Dept.)

In summary I thought that asking the community (easily 2E10 times larger) was the smart move rather than letting an office manager with virtually no IT experience terminate a contract with weak data.

Thanks and Sorry for the Really Long Post
-David Snr CAD Supervisor
 
So your IT dept. is replacing the Dell PSUs with larger EA-500Ds? Which is failing?

Dell often uses Delta PSUs in their builds as I recall, which is one of their strong points.

I recall that a year or two ago Seagate had a lot of problems with some of their drives, but have heard no reports of issues with the newer ones.
 
Try reading the responses in the "Dell" section of this forum, under "Computer Brands", sub-category of "Electronics".

Your manager would be making the right decision if he/she went with HP, over Dell.
 

wigginsd88

Distinguished
Jul 9, 2010
19
0
18,510
So in fact you think that this (IT Replacing the PSU) may have caused the PSU failure? I was not aware that Seagate had these problems. Thank you. I have forwarded the relevent info to the Dept. Manager. Upon looking into this he decided to instead task me with finding an alternative solution for the issue. So I had the IT Dept. Head put the Delta PSUs back, and order Kingston SSDs. The boards now work like a charm. Thanks for the info. If you have any more relevent information, feel free to share.
 

wigginsd88

Distinguished
Jul 9, 2010
19
0
18,510



After a brief review, I have given a green lite to begin negotiations on a contract with HP, but also have suggested hiring a new IT Dept. Head. It would seem that our server lost a drive today. Normally this would be fine, as we have raid configs, but he decided to instead us spanning rather than mirrored. So now he set the office back 2 weeks! :non: He is cleaning out his workspace on Monday. I get to pick the guy's replacement. So I am going to be picking for a good while.

Dell is probably going to try sue us for breach of contract, but if they do, it beats the Department of Energy doing it mostly due to drive failure.
 
Wiggins, the Seagate problems were in their 7200.11 drives. The .12's are by all reports quite reliable.

If you are using standard, consumer grade hard drives, I'd suggest the more reliable enterprise drives. They are about 70% more expensive, but I think the increase in reliability would be worth the additional cost.
 

wigginsd88

Distinguished
Jul 9, 2010
19
0
18,510




Typically, the entire eastern division uses mirroring, thus the it head's firing. Also, as the benefits tend to go, if we can afford our own servers for holding our licenses for our software, we can afford RAID1. But the guy clearly went cheap, I.E. more storage vs. more data security. Though I might add, I e-mailed that link to the ex-IT guy, attaching it to a picture of the damaged drive in question, along with drive failure statistics for drives staying in frequent hi load usage prior to his termination. :pfff:
 

wigginsd88

Distinguished
Jul 9, 2010
19
0
18,510


Yes, but we have decided to go the route of mirrored (RAID1) SSD (Solid State Drive) Arrays for that extra level of comfort. [ch1 RAID1] -> [ch2 RAID1] whereas ch1 indicates a pair of SSDs and -> Indicates a second and separate pair of RAID1 SSDs for additional storage due to the limitations of the SSD technology vs CAD and GIS documents spanning well over 320 GB. The temperature and warranty differences alone bode well. And the MB/s has nearly doubled.
 
This is interesting. I can't claim to be an IT expert, but it seems to me that what you describe is the main function of the job, cost analysis.

I hope this new IT strategy works out for your company.
 


RAID1+0 is different from RAID0+1

RAID1+0 will perform better in this case.

As for justification, the logic is simple. ONE week's lost work in the CAD dept will exceed the cost of a hundred drives! Not even taking into consideration the aftermath or the loss of morale.

Another aspect to consider is that a day's slip on a project can NEVER be made up (catch up). Because, if a day can be made up, then the project could have been delivered a day earlier (completion time) in the first place.
 

wigginsd88

Distinguished
Jul 9, 2010
19
0
18,510



Quite true across the topic, with one exception: its not RAID 1+0 or 0+1, its a C:\cad[edited for security reasons] with a D:\GISDAT[as above edited] where both C and D or for linux users hda0 hda1 ARE mirrored and separate. RAID 1+0 was considered but ultimately discarded as a possibility due to the increased likelihood of failure. However, if you have a different opinion here I always love to learn about this stuff, and don't like to work overtime just to do it again. So by all means if you disagree, feel free to say so, as I mean no disrespect. And again thanks for the help. :D
 

wigginsd88

Distinguished
Jul 9, 2010
19
0
18,510


As I'm sure your aware, that is an issue, and quite a large one at that. We would do this, excepting the sensitive nature of being a environmental geological and surveying company with both civilian aka private sector contracts and government contracts. When corp. is attached to the mailing address, be it USMC or something like Green Power Corp, the contract holders typically prefer we didn't for security purposes. BUT, we do use tape back-ups stored in a locked closet/spare server room on a closed network. It never ceases to amaze me how often the tape drives are actually used. But one might say the if a certain ex-IT head was a little more Tape-drive inclined, then the entire CAD and GIS dept. might have gotten to enjoy this little thing called a weekend, though we seldom do. Though as for the offsite, Division tends to use the crap out of the VPN now that its maintained sufficiently. So I would presume they intend to superimpose themselves in place of the trusty tape drive. But at least its their fault if it screws up.

But I digress, grossly... The problem was ultimately the PSUs, the Seagate HDDs, and an IT head of dept. with relatively no practical knowledge of buyer beware info, or at least no desire thereof. Dell's (TM) name is cleared.
 

wigginsd88

Distinguished
Jul 9, 2010
19
0
18,510


As presumed, the VPN is already bogging down and thanks to an "accidental" VPN encryption lock-out- :) - has been veto-ed at the highest level on a certain 'office-manager-of-the-year-nominated-guy-named-@%#%#&@' we have decided to just be sensible and use the tape drive. When diplomacy fails, reset the password after making a tape back-up, shrugging your shoulders, and oops, you just typed ijhoiuhop for a new encryption key. Now how did THAT happen?