• Content count

  • Joined

  • Last visited

Community Reputation

0 Neutral

About screwtop

  • Rank
  1. screwtop

    LTO tape minimum streaming speeds

    Just a quick update. I've just been testing LTO-2 media in my HP Ultrium 960 (LTO-3) drive, and it will continue to stream happily with no interruptions or shoeshining at 10 MB/s. This could well be slow enough for a typical single disk source, and quite possibly even 100 Mb Ethernet. Note that I'm using mbuffer to buffer and throttle the data on its way through. -- screwtop
  2. screwtop

    LTO tape minimum streaming speeds

    Some useful tips here - thanks! I'm also wondering about the effect of disabling compression on the drive. Even with no compression, this would still leave the (say) 27 MB/s minimum data rate required for streaming, correct? In your experience, have you found that multiplexing complicates the management of backups and archives on tape, or the restoration process? Hmm, it sounds like a dedicated disk or striped array is going to be necessary to get reliable streaming to tape in my environment. I'm still intrigued by the suggestion (in a Fuji white paper, IIRC) of using previous-generation media to "lower the bar" a little on the minimum data rate...
  3. I can't see the original pics either, but if the dips in the sequential transfer rate are in the same place every time and there's definitely nothing else using the drive (use a performance monitor and check that the queue lengths on the drive don't go up when the performance drops), it could be a region of bad sectors that have been remapped. Zero-filling the drive probably won't help much if that's the case, although you should run the vendor diagnostics on the drive. I had a pair of WD1200JB drives that had erratic STR performance, and would sometimes fail the diagnostic test because it was so slow. As for poor access times, check the AAM settings on the drive. I've noticed on some drives that there's a difference between AAM on maximum performance and AAM turned off completely, so try both. HTH -- screwtop
  4. I'm looking at acquiring an LTO-3 tape drive (HP Ultrium 960) and have some concerns about the minimum data rate required for streaming in my environment. I understand the data rate matching (DRM) on the HP 960 allows it to slow down to 27 MB/s, but even without compression this might be too fast for collections of small files being archived on the fly. I understand it's possible to use LTO-2 media in an LTO-3 drive, but will this actually result in a lower minimum data rate? Or will I just have to bite the bullet and buy a dedicated hard drive just for staging the archives prior to streaming to tape?
  5. screwtop

    Deskstar woes due to bad SATA cable

    That's what I'd like to know, too. This cable was installed in the system from new, and had no creases or folds, although there was a minor sideways kink in the cable from it being routed around a SCSI cable. Seemed to work OK initially and then go bad over time. I wonder if the contacts in the cable connectors might be the problem. Anyone know for sure if putting the right-angled connector at the motherboard end is "wrong"? -- screwtop
  6. My home computer has been pretty unstable lately: occasional BSODs under Windows XP, lockups, and spontaneous reboots. A couple of weeks ago Windows started running CHKDSK on my applications volume (on a Hitachi Deskstar, 160 GB, HDS722516VLSA80) on every startup. This happened about half a dozen times, and then the filesystem became unreadable. The drive letter still appeared in Windows, but the volume appeared to be corrupt. I ran some recovery software on it - all the files were still there, but the MFT was apparently corrupt - so I got all the files onto another drive, and then set about testing the drive. The Hitachi drive diagnostics showed no problems with the drive. I reformatted the partition and copied the recovered files back. However, the problems continued. Windows would often reboot during startup, and then have to check the filesystem again. It would boot into safe mode OK, so I started wondering if the nVidia SATA drivers (the drive is on an nForce3 250 Gb motherboard) might be to blame, although I didn't get as far as uninstalling them. I also noticed (after normal Windows bootup - not safe mode) that the drivers had downgraded the controller to PIO mode, so I wondered if maybe the cable might be faulty. So, I replaced the cable with another, and that seems to have fixed it! No lockups, reboots, bluescreens, or DMA downgrades so far... I should have checked the drive's SMART parameters earlier, but I notice now that the Ultra DMA CRC Error Count has a data value of 108. I assume the Hitachi diagnostics would have reported any SMART parameters outside the normal range. FYI, the bad cabling was COMAX E173433-D, 30 AWG, 50 cm length, with a right-angled connector at one end. I'd used the right-angle connector at the motherboard end because it routed under the video card more easily. The new cable uses COPARTNER E119932 26 AWG cable (and, as it happens, quite a bit longer). Noticeably thicker cabling. Thanks for reading - hope this might be useful for others. -- screwtop
  7. 1) Note that the firewire interface will limit the transfer rate of the drive, whereas turning on/up AAM will increase the seek times and therefore the access times. That is, AAM has the potential to slow down the drive in a different way. However, depending on what you're using the drive for, you might see almost no slowdown after enabling AAM. For the kind of desktop-style use you describe, AAM is probably fine; for server use, you'd want to turn AAM off. 2) I'm not sure about the power management options for Hitachi drives - but it occurs to me that AAM might have an effect on power usage as well as seek times and noise. I'm sure someone else here can clue us in on this one. Read look-ahead cache is almost certainly enabled by default, and almost certainly should be left enabled. Most normal (desktop) kinds of disk activity are significantly speeded up by read-ahead, and you'll notice a big drop in STR benchmarks (FWIW) if you turn it off! Hope this's definitely worth making sure things are set up properly before you move the drive to the Firewire connection. -- screwtop
  8. IIRC, the recommendation about leaving drives running was mainly to do with temperature extremes and mechanical wear. A drive without forced air cooling could be subjected to a 20 degC change in just a few minutes on a cold morning. I know for shipping the maximum temperature gradient specified by the manufacturers is about 20 degC per hour! You could argue that with a good cooling airflow, the drive temperature won't go nearly so high above ambient, and the temperature gradient would be lower, so you could safely spin the drive up and down more often as required. Either way, cooling is a winner! You could probably make the case for powering up the drives and leaving them to reach thermal equilibrium before actually using them. I seem to remember reading that the "leave them spinning" recommendation was more applicable to older drive designs - perhaps in the days of stepper head motors and no thermal calibration. But I do remember seeing some info on some recent desktop drives that recommended against using them 24x7. -- screwtop
  9. Hey, what great timing - I was just looking for a tool for monitoring and analysing disk accesses on my swap drive, and lamenting the fact that the IPEAK SPT demo has a limit of only 5 launches. You can specify the drive to monitor and sampling interval (in ms, I think) on the command line, e.g. HDiskPerf.exe -Disk:0 -Delay:1000 (There are additional command line options for partition and log file, but they didn't seem to work for me.) Unfortunately, it won't let you run more than one instance at a time, e.g. for monitoring more than 1 drive at a time. Come to think of it, a queue depth gauge would be nice, too...and different colours for read and write operations... Nice find! -- screwtop
  10. Yep, Western Digital tech support agreed - both my drives should be sent back for replacement. I was surprised at how much a zero fill seemed to help things, though (even though the drives were still having problems). I wondered if the lower ambient temperatures here (New Zealand) compared to the drives' place of manufacture (Malaysia) could cause problems. Cheers, screwtop
  11. screwtop

    This is amazing

    Urghh, motion sickness... Pretty kewl, tho...
  12. Hey all, I'm having what appears to be the same issue with my two WD1200JB drives. I set them up as a RAID 0 volume on a HighPoint HPT370A PCI controller card when I first got them, and was never really impressed at the performance. In fact, it sucked: I've now had a chance to move the data off the array and test the drives individually to see what was going on. Both drives showed highly erratic transfer rates, with very low minimum and average transfer rates, and access times were a couple of ms off spec. The striped configuration probably exaggerated the problem. As suggested by Western Digital tech support, I performed a zero fill of the drives to see if that would help. Drive #1: Zero Fill: OK Quick Test: "Completed with read element failure", error code 0007, checkpoint 65 Second Zero Fill: OK Quick Test: error 0007 again Extended Test: bad sectors (2 of them) found and repaired Quick Test: OK Here's a before and after comparison using HD Tune: Before zero fill: After zero fill: The STR performance now looks normal, and access times are now on-spec. The question is, would you trust this drive now, or should I get an RMA, given that the drive continued to fail the Quick Test after the first Zero Fill? Drive #2: Zero Fill: Many errors from sector 86,443,008 onwards; drive stops responding ("Drive Not Ready", error code 0104); HDD Activity LED remains on Extended Test (after cold boot): "Too many bad sectors detected" Second Zero Fill: More errors in same area; drive becomes unresponsive This one's definitely going back! Before (partially successful) zero fill: After: Notice how the STR now looks normal over the part of the drive that was able to be zero filled. FYI, both drives were made in Malaysia, 23 March 2003, serial numbers in the range WD-WMA8C371xxxx. -- screwtop
  13. Thanks, MaxtorSCSI, it's good to get the definitive word on that! I did find that changes to the minimum prefetch were also not retained by the drive, but now I know not to worry about it. BTW, I should mention that I am very pleased with the performance of this drive - I just wondered if I could wring anything more out of it! -- screwtop
  14. Well, it pays to RTFM, which states that changes to the Number of Cache Segments parameter are ignored by the drive. I'm still interested in tweaking the minimum number of blocks to prefetch, which defaults to 0 and is supposed to be changeable. I'll try again later tonight and see how I go... -- screwtop
  15. I'm trying to configure my Atlas 15k drive for better performance in its current non-server installation. In particular, I'd like to try changing the number of cache segments (currently 21) and the readahead cache behaviour to see how that affects things. Would lowering the number of segments be a good idea for single-user disk accesses? I've been using Bart's SCSITOOL for MS-DOS (great tool) to try to make these changes, but the Atlas seems to be ignoring the changes to the mode pages. Are there any other tools for changing SCSI mode page parameters (preferably Windows XP or DOS) that anyone can recommend? BTW, handy thread on SCSI mode page parameters here (e_dawg's posting halfway down the page). Cheers, screwtop