• Content Count

  • Joined

  • Last visited

Everything posted by mejv

  1. mejv

    Large Drives Compatible with 9500S-12

    Sorry I did not precise: The 32bit LBA is for the drive (HDD) not for the array. The array would support 64 bit LBA. It has from day 1... Regards, MEJV
  2. mejv

    Large Drives Compatible with 9500S-12

    For what I know, the 9500S is a 1.5Gbps SATA Raid controller that does not know anything about 4kB sectors. At the time it last got a FW release, 512 bytes sectors were the only ones available. As for capacity, the 32 bit LBA (2TB of 1024 bytes per K) is the MAX supported. Hope this helps! Regards, MEJV
  3. Could it be that the source drive is fragmented and the HDD has to seek all over the place? MEJV
  4. mejv

    3ware raid keeps rebuilding!

    Send a more complete log... Maybe a drive had a pbm... or the controller... If there is a way to gather the FW log as well... Regards, MEJV
  5. The implementation of the caching algorithm is specific to the vendor. It would depend also if the data access pattern from the host: purely random, or not: if the controller FW think it is possible to improve the performance by reading ahead (100% sequential) or not to prevent wasteful overhead (100% random)... Regards, MEJV
  6. As previously stated, if the host requests 4KB, the RAID controller would provide 4KB only. If the controller need to get more from the drive(s), that is up to the implementation of the RAID controller. If, instead of a raid controller you have a software RAID, the software RAID would decide how it need to get the 4KB (could be spanned over 2 drives)... Maybe you are getting confused on the various layers involved and do not ask the right question... Simplistic view of the layers: Application->OS ->Driver->RAID controller->disk drives 4KB req. -> 4KB->4KB ->4KB -> 4KB to 1 drive or combination over 2 drives for 4KB total. MEJV
  7. I am not too sure of what you mean here, but the 64KB strip is how much data is stored on one drive in LBA sequence before switching to the next drive. Your 4 KB example could fall onto a single strip, so one drive or split onto 2 strip, so over 2 drives... Regards, MEJV
  8. For the application, the request being 4KB, the Raid controller would provide 4KB as requested, no more, no less, since the raid controller simulate a HDD to the application (OS). The application/OS is not aware of the internals. Hope this helps! Regards, MEJV
  9. It is not even Beta but a released code: Good luck! MEJV
  10. Firmware version was released over 2 years ago... There were no over 2TB drives back then... The latest beta release since last year support > 2TB drives... Good luck! MEJV
  11. Hi there, could you provide a bit more info: Firmware version, Drive model and reported capacity, raid type... Thanks! MEJV
  12. I'd try to reduced the drive link speed to 3Gbps as the drives are 6Gbps. The 9690SA is only a SAS 3Gbps... good luck! MEJV
  13. These should be a jumper to prevent the drive from using its fastest link speed. Try to set it, it might work... Let us know... Good luck! MEJV
  14. mejv

    RAID dropping drives on reboot

    Hi Halo, The 30 minutes sounds like a sleep timer (OS?), drive or JBOD setup timer to spindown the drive, leading to some unexpected drive spower-up delay leading to the dropping of the drive late to come up. I'd check for any sleep or staggered spin values on the controller/expander... Good luck! MEJV
  15. I have used xxclone with success in the past (windows XP). It is freeware! MEJV
  16. mejv

    The Hard Drive Giveaway Thread!

    I'd be happy to get some 2TBs... any brand... Thanks! MEJV
  17. mejv

    Incorrect Spindle Speed Reported

    Is it 65k or 65535? 65535 == 0xffff which sometimes means Not Valid... MEJV
  18. mejv

    Moving 3Ware 9750-4i4e between servers

    Hi Halo-wings! Moving the existing Perc4 controller and its Raid cage to another system should work without initialization, assuming OS driver exists and the Raid file system is supported(unless HW incompatibility issues). In any case, you should have a current backup of the data no matter what. Regards, MEJV
  19. For the SCSI commands B7h (Request Defect Data), there is nothing to worry about (Recovered Error, no retries): the controller is checking the defect data on the drives and the format requested by the controller is not supported, so the drives sends a different format of the data and notifies the controller that the format returned is not the one requested. It is up to the controller to implement the proper format decoding (just a few cases) so most controllers should have it implemented. As for the other error. it is on a Log Sense command where the controller is asking for some information (page 1Fh) and the drives do not support it and fails the command with the SCSI standard way: PAGE CODE Field The PAGE CODE field specifies which log page of data is being requested. If the log page code is reserved or not implemented, the command shall be terminated with CHECK CONDITION status, with the sense key set to ILLEGAL REQUEST, and the additional sense code set to INVALID FIELD IN CDB. Regards, MEJV
  20. A 3ware connection indicated that the >2TB changes was implemented for 9650/9690 after the 9.5.3 release, in the beta code due to lack of >2TB drive availability... Any one tried the beta code with >2 TB drives? MEJV
  21. I thought the latest beta firmware from 3ware (v4.10.00.019 November 2010) had some notes about >2TB drive support... Anyone tried it? MEJV
  22. This whole verbiage answers exactly why the HDD manufacturer delayed creating new changes to pass beyond 2TB size dictated by 32bit limitations. The development cost, time, testing, OEM qualifying of the new line of HDD (<2TB and >2TB) is not something they'll do for your fun. They'd do it if they must do it to keep market share and volume. As far as I know, the sweet spot for today is not > 2TB, but the trend is closing in, so that's a good reason why changes are coming now rather than 2 years ago.
  23. "Are you suggesting that hard drives are using 32-bit controller chips to save money and are limited to 2TB, or that 32-bit CPUs are being reused on PCs to save money and cannot handle >2TB drives? I've already proven the second one false and can prove the first one to be rubbish as well." All I am saying is that HDD companies do not want to spend development time and money if they don't have to, so they reused/leveraged as much as possible from previous designs. As far as firmware or software, 32bit is not much of a limiting factor besides the investment in the code where 32bit words was a fast path for computation speed and structures alignment. Changing these is a change that requires careful testing and qualifying for OEM, therefore expensive.
  24. This is the start of the misleading information stating that 512 bytes sectors (not 512 bits...) does not have limitations... From a software point of view, there are no limitations... from a hardware point of view, like a HDD, where price is the most critical issue, 32bit processors and hardware were used and re-used as much as possible for many generations and using 32bit words and registers was fine until its limit was reached. 4 KByte sectors would somewhat postpone this 32bit limit for a couple of years, not for ever, but 4KByte is a common size use for 'clusters' of some kind at OS level... so having a 4 KBytes sector makes sense... Getting the industry to switch to that won't be immediate for downward compatibility reasons...
  25. Again, you don't have any idea of reality, why does microsoft 32 bit OS does not support more than 4 GB RAM and over 2 TB HDDs? After all, according to you, it is just a software change... Yet after so many years... But what do you know? If you know something we don't, why don't you explain to us how to do it? Ah, and don't forget to justify why controller companies should invest to update the firmware on older controllers while you're at it! Thanks!