GrandAdmiralThrawn

Member
  • Content Count

    7
  • Joined

  • Last visited

Community Reputation

0 Neutral

About GrandAdmiralThrawn

  • Rank
    Member
  1. Greetings! I would have a question about our HP MSA2312fc SAN system. We have HP #480938-001 (HP-branded Seagate ST3300656SS) disks in that SAN, and I would like to replace two drives which are slowly becoming bad. There is no more warranty or carepack. Those HDDs are essentially Seagate Cheetahs, 15k.6 300GB SAS. So, that original HP part is unbelievably expensive, hard to get and the regular Seagate ST3300656SS also unavailable. I also asked HP themselves and they refused to send me the disks alone, they only do it if we pay their technician to replace the disks on-site for lots of money. Ridiculous.. As if I couldn't do that myself! So I thought I would like to replace the slowly dying drives with ST3300657SS (Cheetah 15k.7), the direct successor of the drives we have, those drives are still available. Size is ok, I compared both disks specifications and they're exactly of the same size (same sector count). Now the question: Does a drive NEED to use HP firmware to work in an MSA2312, or can I buy any off-the-shelf Seagate SAS drive and use it for my RAID arrays in there, as spares / for rebuilds? Has anyone ever tried this? Thanks!
  2. GrandAdmiralThrawn

    How to detect speed of 64-Bit/66MHz PCI (and PCI-X)

    Ok, thanks, didn't know that tool before. I'll give it a shot.
  3. Greetings! I hope this is the correct subforum for my question, I wasn't exactly sure where to post this. So, I'm looking for a reliable way to detect in what mode of operation a given PCI card and/or bus is (mostly for determining PCI bottlenecks for RAID controllers). For PCI Express this can be determined quite easily (1.1/2.0/3.0 mode and number of active lanes). But I'm looking for regular parallel PCI. So I would like to have a software tool, that can tell me "Card XYZ on bus 0, slot 0 is running 32-Bit 33MHz" or "Card ABC on bus 1, slot 1 is running 64-Bit 66MHz". I know that on Unix "lspci" can do it (both reading of theoretical capabilities of a device and also the actual mode of operation), by just running lspci -vv as root. But I'm looking for a tool that can do that kind of PCI scan reliably under Windows. Sandra/Everest have not proven to be realiable really. Any ideas? Thanks
  4. GrandAdmiralThrawn

    Storage Review Site Update

    I recently bought a Sony Vaio TT subnotebook and upgraded it to a nice Intel X18-M SSD from a crappy 1.8" mechanical drive. I think, some decent SSD coverage for both (ultra)mobile and desktop markets would be a good thing. But I'm also a mass storage fanatic, running a RAID-6 using a 3ware 9650-SE Controller and 8 x 1TB Hitachi drives. So personally I would be interested in RAID reviews even more than SSDs. Which controllers "really" have the fastest XOR processors? PPC/IOP33x/IOP34x etc. How good is R/W caching working on different controllers? How to they scale in performance in RAID-5/6 when adding more drives? Plus some real world benchmarks, from Desktop User over Workstation up to File-/Webserver usage patterns. Stuff like that. Also it would be nice to show the actual benefits (both in reliability and performance) of drives, that are explicitly targeted at the RAID market, like the Western Digital RE4, when actually being run in RAID. I never get it, why websites test this drive (and also others like the Hitachi Ultrastar) in Single-Drive configurations, just doesn't make much sense I would say. I'm still waiting for a website to give me a review of Hitachi Ultrastar and WD RE4 compared in RAID-5/6, maybe together with a Barracuda XT and a RE4 Green Power. I would definitely be excited about that.
  5. GrandAdmiralThrawn

    3ware 9650SE-8LPML major write problems!

    Sorry for the double post, i had problems posting yesterday, my post just didn't appear, so i tried a second time. Not so good obviously..
  6. GrandAdmiralThrawn

    3ware 9650SE-8LPML major write problems!

    Greetings! I'm currently using the same controller, with 8 Hitachi 7K1000 1TB drives attached. I have contacted Intel support about the Memory Write and Invalidate (MWINV) Bit, and this bit has been removed from PCI Express specifications since the beginning. I can confirm this bit being hardwired to "0" on i975X (PCIe 1.0a), because i tried to set it to "1" using PCItree and WPCREdit. To no avail. I do however get sustained sequential writes around 50MB/s using all 8 drives in one RAID-6 array, also using SCSIPort, not StorPort (WinXP Pro SP2). So maybe it's a Firmware or driver thing? My versions: Firmware FE9X 3.08.00.004 Driver 3.00.02.090 BIOS BE9X 3.08.00.002 Boot Loader BL9X 3.05.00.002 Now quoting Intel Support:
  7. GrandAdmiralThrawn

    3ware 9650SE-8LPML major write problems!

    I'm using the very same controller, with 8 Hitachi 7K1000 drives attached. I have asked Intel Support about the Memory Write and Invalidate (MWINV) Bit, and they told me, that MWINV has been completely removed from PCI Express specifications (at least with PCIe 1.0a). As far as I know today, MWINV is NOT a PCIe feature at all, but only usable on parallel PCI buses. You can't even switch on that bit using PCItree or WPCREdit. It's hardwired to "0". However, I'm currently using SCSIPort (WinXP Pro SP2) and the 7K1000 disks in one big RAID-6 array, getting around 50MB/s sustained sequential write speeds, benchmarked with ATTO Bench. Intel Support quoted: