• Content Count

  • Joined

  • Last visited

Community Reputation

0 Neutral

About tdiroff

  • Rank
  1. Hi folks. I was wondering if any of you learned folks have recently explored the viability of using ram disks in order to maximize performance in a server environment. If so, can you share your thoughts and any insights? Thank you Tom D
  2. Hi Yes, it (Ghost 8.2) does allow you grab an image without having to load anything into Windows. Tom D
  3. tdiroff

    Ghost & Trueimage images

    Hi I have this happen to me often and the solution generally is to do Windows repair install on top of the new image to repair whatever is wrong. Good Luck Tom D
  4. Hi We use Ghost 8.2 a lot and it seems to work pretty well even though it will complain about dynamic disks, it can work. We also use it with iSCSI disks with very good success. We are also taking a look at image for windows from terabyte for a good low cost Windows option Good Luck Tom D
  5. HI I would turn off all onboard stuff that I didn't need to boot like nic's etc. Also play around with onboard video vs add-in video. We had problems with new Broadcom Nic's and ATI ES1000 video, the video driver must be loaded first or the nic drivers would blow it up. Tom D
  6. Hi I have been working a little with Image for Windows and have been pretty impressed given the price. I can do a live image for about $40/server.... not bad! Tom D
  7. tdiroff

    Cloning unknown file system

    Frank, I never saw that patch, do you still have it? It would be interesting to understand. On a side note I am not sure how a PE copy of Windows (Barts PE) works in this area but my guess is that it still likes to add signatures or metadata. Does anyone know? My other thought is that Windows can never be used for forensic tasks, even on other Windows drives, for exactly this reason. Tom D
  8. tdiroff

    Cloning unknown file system

    Hi I would tend to stay away from mounting it in an Windows system for imaging as Windows tends to want put signatures on the drives and that may affect it's operability. If this happens how would you recover in this case? I think that Franks's Linux and dd approach is one of the best to use Tom D
  9. tdiroff

    Decent Matrix RAID article

    Nice catch Frank! Thanks!
  10. Hi We are looking at a system that could support SAS and SATA drives. It is my understanding that if you have a controller that supports both SAS and SATA drives, the SATA speed will be 50% of the SAS speed because of the design of the SAS/SATA drive connector. So if one wanted top SATA speed (3.0Gb/s), one would have to have a SATA only system (or have SAS drives capable of 6gb) Am I correct in my thinking? Thank you Tom D
  11. tdiroff

    Benchmarking ideal stripe size

    Hi Frank If you ever find that web site please let us know, I would love to read up on it! Thanks Tom D
  12. Hi Perhaps, but if that many CPU's had to be replaced it would suggest (to me ) that the nodes under study were of questionable quality and cast some doubt on the entire result set. We really don't know, but it raised a red flag for me Tom D
  13. I read the fisrt study about a week ago and found some of the results to be so far away from personal experience that that I could not trust the rest. (In particular 40% of node hardware failure due to defective cpu.... I find it extremely rare to have to replace a cpu to fix a hardware problem)
  14. Hi Awesome document! Thank you!! I would love to see the breakdown of mfg and model in this study! Tom D
  15. Hi Yeah, you are right too many of my questions are too general but I guess even I narrowed it down to "are Hitachi/IBM drives made today better than Hitachi/IBM drives made 4 years ago" it would be difficult to answer but nevertheless an answer I would love to know. This also runs in to the problem of what does "better" mean and while there are many measure I guess I would be happy with less failures/1000. (happy until I get that number and then I would want something else.....) The mention of MTBF (or AFR) is interesting because some will claim that it is a false number and has not real decision making value. I would love to see defect numbers versus stated MTBF versus manufacturer to see who is embellishing more than others. Yeah, SAS etc are connects but they will also generally define the class of drive (I don't think I seen too many Enterprise PATA drives). In specific cases that we face today, if we but a SAS drive chassis we can populate with high cost SAS drives or lower cost SATA's. If the SAS's are truly more reliable (less failures/1000) perhaps the higher cost would be justified. The bathtub curve is one that is interesting and the more I think about it I think that it really is a composite curve where firmware problems would ONLY show up close to t=0 and mechanical wear at t+x where x would tend to vary from mfg to mfg. If so, with proper data, we should be able to see which mfg does a better job with various components of the drive (firmware testing, mechanical stability, etc etc ) Tom D