Page 2 of 3 FirstFirst 1 2 3 LastLast
Results 21 to 40 of 49

Thread: When Sleeping, MacPro Turns Itself Off

  1. #21
    Join Date
    Sep 2003
    Posts
    310

    Default DiskWarrior vs. TechTool

    In listing bad blocks, apparently DiskWarrior simply reports what the drive's internal diagnostics say, whereas TechTool supposedly tests every block on the disk. In the case of my boot drive, DiskWarrior says that about 60 attempts have been made to automatically map out bad blocks, exceeding the entire reserve of about 30 spare blocks. If that's true, there are about 30 bad blocks lurking on the disk that TechTool is unable to find. Useless!

    Replacing the drive would likely be the best course, but with me already skimping nickels to buy some more MacGurus RAM, I'm going to wipe the drive for now. I'll keep a clean disk image of the fresh installation, to make replacing the drive easy when the time comes.

  2. #22
    Join Date
    Aug 2001
    Location
    Grangeville, ID USA
    Posts
    9,119

    Default

    Might download and try the trial SoftRAIDLite to see what it reports on the drive.

    R
    molṑn labe'
    "I am a mortal enemy to arbitrary government and unlimited power. I am naturally very jealous for the rights and liberties of my country, and the least encroachment of those invaluable privileges is apt to make my blood boil."
--Ben Franklin

  3. #23
    Join Date
    Sep 2003
    Posts
    310

    Default Sobering SoftRaid Lite

    Thank you for the tip about SoftRaid Lite. I am always amazed at the arcane knowledge MacGurus has about Macs. Great utility! It answers a lot, and also raises some questions.

    The sobering news is that three of the tower's four drives report "disk failure predicted," although two of those verify with no errors. The boot drive, a 500GB Hitachi DeskStar from MacGurus, has 634 reallocated sectors and 7 unreliable sectors, and 6,672 hours of use! Its twin has 1 reallocated sector, and 4,526 hours of use. They were installed together, yet clearly the boot drive has spent far more time spinning. The most astonishing disk is an old, old one with 22,611 hours! It has 466 reallocated sectors, and 14 unreliable sectors.

    Amazing figures, but what is their significance? Are hundreds of bad sectors among millions considered bad or trivial? Does this mean it's time to reformat, to map out the bad blocks; or, when some blocks fail, will other surely follow, so that the time has come to replace the drives before a sudden, catastrophic failure?

    Verifying only reads the drive; it does not write to it. Can a drive that reads well independently have trouble writing?

    SoftRaid "certifies" a disk by wiping it, and then both reading and writing to it repeatedly. Is that truly any more effective than simply securely erasing a disk multiple times?

  4. #24
    Join Date
    Sep 2003
    Posts
    310

    Default P.S. -- SoftRaid Reports Boot Drive "Hung"

    In three attempts, SoftRaid was unable to complete verifying the boot drive, because the drive "hung." Whatever that means, it can't be good. Does that mark this drive as an especially bad candidate for wiping and reformatting?

  5. #25
    Join Date
    Aug 2001
    Location
    Grangeville, ID USA
    Posts
    9,119

    Default

    Realizing this might not be an answer you want to hear - those 500 GB drives are now 8 years old. I have always considered desktop drives to be a 3 year investment. I talk about the advantage of Enterprise drives being their reliability over a 5-6 year period. You have gotten fantastic service from those drives.

    I do not ever multiple erase and write zeroes to drives. They should map out the bad blocks with a single pass of zeroes. Doing it multiple times seems like it is just extra wear and tear for no advantage.

    I do not know how accurate SMART reports are. In the 'old' days (2005-6) there were reports and studies showing SMART being somewhere in the high twenties percentage accurate in predicting impending doom. Franky, if you study what is being written now, people are all over the place on how good the predictions are. Kind of like predicting your own death - it is inevitable, eventually. But I can't tell you how many people are telling me that SoftRAID predicted multiple impending drive failures. It is a good check to see what SMART is reporting, but is it true? I'll see if Mark James will post his thoughts. He's the expert here.
    molṑn labe'
    "I am a mortal enemy to arbitrary government and unlimited power. I am naturally very jealous for the rights and liberties of my country, and the least encroachment of those invaluable privileges is apt to make my blood boil."
--Ben Franklin

  6. #26
    Join Date
    Sep 2003
    Posts
    310

    Default Time for New Drives; Happy About Eight Years of Reliable Service

    I'm curious to hear what Mark James may say, too, but the writing's on the wall! Time for new drives. I'm actually not unhappy to hear it, although I am freaked out to hear that the failing drives are eight years old. That suggests that I may be, too; I had no idea I had had those drives for that long. One goes to the Gurus to get the best, for the most reliable, lowest cost of service over time, and they always deliver. As I've said, I am a loyal, if only occasional customer. Amazing the stuff I have bought here.

    This requires a change of plans. I was about to (finally!) order more RAM for my laptop, but am thinking I'd better verify the laptop's drive with SoftRaid in case it's failing, and then buy strategically. It would be beyond budget now to replace all three failing drives in my tower, and possibly the one in my laptop. Looking ahead, what would be the one smart buy now, to get the tower ready to have, eventually, one drive for its OS, one for its data, and one giant drive for backup? Or, are drives so fast now that I could be happy with just one drive for OS and data, given that I am not crunching video or, honestly, currently doing any taxing work with this Mac?

    I've learned a lot watching Activity Monitor, since first thinking of adding RAM to the laptop. Surprisingly, unless a Web page is badly written, it is difficult to overwhelm the processor. I can open a million Web pages and too many applications, and, just to keep them open, the processor scarcely runs. This is especially true in the eight-core tower, yet also holds for the two-core laptop. When the computers slow, it is clearly due to paging to disk. So, yes; more memory would cure that, but a new drive or drives would keep my computers running. Better to be reliable than somewhat faster!

  7. #27
    Join Date
    Aug 2001
    Location
    Grangeville, ID USA
    Posts
    9,119

    Default From Mark James...

    Mark James' reply:

    SMART is typically just a yes/no check. (The SMART test in Disk Utility as an example.) This binary SMART test is considered by industry experts to be fairly useless. If a device fails the SMART yes/no test, it is already failed.

    There are dozens of parameters that are available to a SMART query. Many of the SMART parameters appear important, but what do they really mean? Which ones indicate "poor health"?


    The first large scale test of SMART's usefulness was a Google study of 100,000 disks, where the question was essentially posed, "out of all the SMART parameters, which ones are most predictive of future failure"?


    The results were fascinating, as all the SMART parameters were compared statistically for correlations with disks that subsequently failed.


    Most of the "common sense" failure indicators (i.e. power on hours, number of power cycles, peak temperature, sustained temperatures, and even whether the drives were used for light duty or heavy duty cycles) did not factor in the results. Surprisingly, it did not matter how heavily or lightly a disk has been used when you want to predict whether a disk may be prone to failure.


    The parameters that were predictive of future failure were all related to reallocated sectors: Whether there are any reallocated sectors, pending reallocated sectors, failed reallocated sectors and unreliable sectors.


    Disks where any of these parameters were >0 were found to be 20 to 60 times more likely to fail in the next couple months. Not surprisingly, when any of these specific parameters changed, it was highly predictive of an imminent failure.


    BackBlaze, a cloud storage/backup vendor, has been publishing their stats, including disk manufactures, in the past few years. Their blogs are fascinating reading for those interested in this subject. BackBlaze has independently validated the google study findings.


    BackBlaze also found disk errors had predictive failure, although their disks are in standardized, controlled pod environments. In their hardware, and because they can the ability to track disk errors, a disk error becomes an important event. Disks which generated errors was found likely to fail. BackBlaze apparently implemented a policy where if a disk had an error, it was on probation. A second error and a disk is replaced, regardless of any other factor.


    In consumer hardware, disk errors are often caused by temporary failures in hardware, so disk errors cannot be considered as predictive. However, SoftRAID does track disk errors, and we recommend a user replace a disk which has an error. That is why SoftRAID posts the disk errors in the disk tiles in the GUI. Of course, because this is consumer quality hardware, many disk errors can be explained by other factors, such as loose cables, power fluctuations, RAM, PCI cards and more. So the user needs to decide whether a disk that experienced a disk error was caused by another factor or not. Without a good reason, if this is mission critical data, then replace the disk.


    SoftRAID uses this same set of parameters involving reallocated and unreliable sectors for predictive failure reporting. We do not consider any other parameter as a valid predictor of future failure.


    Anecdotally, we get dozens of users reporting that disks reported as "predicted to fail" have indeed failed on the user.


    SoftRAID also displays power on hours, as while it is not a statistical predictor of failure, it certainly is an indicator of the need to replace disks that are past their expected lifespan. Generally, in a server environment, we suggest disks that reach 20-25,000 hours should be replaced.


    The other thing I should bring up is Certification. Over the years, we have found the SoftRAID feature to "certify disks" extremely valuable. Many users have told us that they purchased new disks which failed certification. Commercial disks are sold untested, so certification is really the first time a disk has been exercised. (There are a couple testimonials on our web site from users who had disks fail certification.)


    The reason this is more valuable than "verify disk" or "formatting" a disk is this: A certify writes a long pattern to the disk, then reads it back. This is done (recommended) three times, and the last pass is all zeros. A simple verify disk only reads the pre-existing zeros on the disk, and is not a test of whether each sector can be written to and read back.


    A certification can take several days, but it is invaluable.




    Mark James
    mjames @ SoftRAID.com

    molṑn labe'
    "I am a mortal enemy to arbitrary government and unlimited power. I am naturally very jealous for the rights and liberties of my country, and the least encroachment of those invaluable privileges is apt to make my blood boil."
--Ben Franklin

  8. #28
    Join Date
    Sep 2003
    Posts
    310

    Default Time for a New Disk

    Thanks, Mark!

    I have many hundreds of reallocated sectors, and 14 unreliable sectors. Time to replace the disk! I'll license my SoftRAID trial, too. I may never actually build a RAID, but what a great diagnostic. Maybe I'll eventually create a mirror as well as a backup.

    Ricks, I await your recommendation on a new disk for my MacPro3,1, early 2008, or should that be off the forum?

    THOUSAND THANKS.

  9. #29
    Join Date
    Aug 2001
    Location
    Grangeville, ID USA
    Posts
    9,119

    Default

    2 TB Seagate. Dazzling fast. And price is such that it makes more sense than a 1TB drive. We must have a dozen of them in the MacGurus office, churning away year after year.
    molṑn labe'
    "I am a mortal enemy to arbitrary government and unlimited power. I am naturally very jealous for the rights and liberties of my country, and the least encroachment of those invaluable privileges is apt to make my blood boil."
--Ben Franklin

  10. #30
    Join Date
    Sep 2003
    Posts
    310

    Default Best Use of Two Drives

    I need two drives, because my backup drive is failing also. The Tech Guides recommend a minimum of three drives: OS/applications, data, and backup. What's the best way to configure two? Would it be better to use one for everything (OS, applications, data) and the other only for backup; or, would the gain in speed make it worthwhile to partition both drives, so that OS and applications could be on the active volume of one drive, with data on the active volume of the other drive, and then backups for each of these active volumes could appear in the backup partition of the opposite drive?

    I'm amazed at the price of these fast 2TB drives. They cost less than junk used to cost!

  11. #31
    Join Date
    Sep 2003
    Posts
    310

    Default Thank you for alerting me to the problem!

    Perhaps playing with SoftRAID has made me more aware, or maybe my system is truly failing more rapidly. Lots of weird, glitchy behavior now. It's shut down, until I can move my data onto new disks.

    Are the 3TB Seagates as fast and reliable as the 2TB? I have many ancient, external hard drives that would be better consolidated onto newer hardware.

    I've decided to keep it simple for now. One drive for everything, with a second partitioned for two backups: One SuperDuper bootable clone, in case the primary drive fails, and one Time Machine, in case I make a mistake while consolidating my files.

  12. #32
    Join Date
    Aug 2001
    Location
    Grangeville, ID USA
    Posts
    9,119

    Default

    I personally always use a pair of drives with 2 partitions. I boot off partition 1 on drive 1 and back it up to partition 2 of drive 2. Then I put data on partition 1 of drive 2 and back that up to partition 2 of drive 1. That gives me twice the performance for the exact same dollar.

    Rick
    molṑn labe'
    "I am a mortal enemy to arbitrary government and unlimited power. I am naturally very jealous for the rights and liberties of my country, and the least encroachment of those invaluable privileges is apt to make my blood boil."
--Ben Franklin

  13. #33
    Join Date
    Sep 2003
    Posts
    310

    Default Must Buy a Third Drive, for my Laptop

    Apologies for vanishing for a week after all your help, without yet buying. Left town to take care of ailing relatives. Now I'm back, so it's time to buy.

    Ricks, I'll buy the two drives for my MacPro, and configure them exactly as you suggest. Great plan, thank you!

    I need to also buy a new drive for my MacBook Pro, because SoftRAID shows its hard drive is failing, too. Is there any reason that the Seagate Hybrid's combination of speed and storage would not make it the drive to get? Is a pure SSD appreciably faster? I'm not worried about battery life and can't say I need more than 250GB, yet it never hurts to have spare storage at hand.

  14. #34
    Join Date
    Aug 2001
    Location
    Grangeville, ID USA
    Posts
    9,119

    Default

    An SSD really shines in a laptop. Since the laptop only has the one drive it has to do everything.

    I am running a Hybrid now and it is only marginally better than a regular spinning drive. The bugs were worked out, so it has been spectacularly reliable. I needed the larger capacity and couldn't justify a big SSD in my MacBook Pro.
    molṑn labe'
    "I am a mortal enemy to arbitrary government and unlimited power. I am naturally very jealous for the rights and liberties of my country, and the least encroachment of those invaluable privileges is apt to make my blood boil."
--Ben Franklin

  15. #35
    Join Date
    Sep 2003
    Posts
    310

    Default Questions for New Hard Drives

    Progress! The SSD arrived, thank you, and was a piece o'cake to install into the MacBook Pro using SuperDuper to clone the old drive's Mac OS X 10.6.8 to the smaller of the SSD's two partitions. The results are ASTONISHING. Everything runs sprightly now. Just amazing. Thank you! I'm still saving my nickels to add more RAM, to see how far it would push the performance, but everything is restored to good usability.

    Now, I want to install the latest Mac OS X on the MacBook's larger partition. Because I need to keep the older OS to run legacy software, I can not update 10.6.8 to 10.7; so, using the currently working partition, I can not install the firmware that would allow me to install El Capitan. Is there an easier way around this than installing on the larger partition whatever OS came on the DVD with the laptop, and then upgrading? I was hoping for a single, clean install, but maybe that does not matter so much today as it used to.

    The two Seagate 2TB drives are currently being SoftRAID certified in my MacPro, provided only that the failing Lion boot drive lasts the 15 hours needed. Something is surely wrong, because the icon for Google Chrome still appears in my dock, but with a question mark superimposed, with the application itself nowhere to be seen. It's not in the Applications folder, and Spotlight can't find it. Is there any problem SoftRAID certifying a new drive from a failing drive; or, so long as the process finishes, can I trust it to be correct?

    I love the "MacGurus tested" stickers sealing the new drives. How do you test them? Do they not need to be SoftRAID certified?

    Thank you for always carrying only what is best to buy. Makes shopping so much easier. The Gurus have always delivered!

  16. #36
    Join Date
    Feb 2001
    Location
    on the landline, Mr. Smith
    Posts
    7,784

    Default

    Hmmm.....maybe Rick knows the answer:

    Firmware -

    It gets installed on a ROM (chip) on the logic board. Yes, it looks like a typical software update, but it is not. The update is actually flashing (rewriting) firmware instructions on the logic board, nothing to do with the HD/SSD.

    The tricky part: You cannot downgrade firmware on a Mac. Once you run the update, it can not be un-run. Normally there is no issue; the update fixes bugs, or security issues, enables features/hardware changes, or other very low level instructions.

    It may be possible that after you run a firmware update, you may not be able to boot an old OS properly again.

    I don't think this is the case.......but you need to be sure be for you run it.
    "Imagine if every Thursday your shoes exploded if you tied them the usual way. This happens to us all the time with computers, and nobody thinks of complaining." -- Jef Raskin

  17. #37
    Join Date
    Sep 2003
    Posts
    310

    Default Sizing Boot Partitions for Unexpectedly Large Swap Files

    Is there a best way to choose the sizes of partitions on hard drives? Is some amount of disk space enough to allow essentially everything to remain open?

    I need to re-partition my new Crucial SSD, to enlarge the 13GB of free space (at startup) on the Mac OS X 10.6.8 boot volume. With this new SSD, the MacBookPro7,1 had been performing astonishingly well, until today, when it posted an out-of-disk-space message. Having filled all 13GB with swap files, it refused to work until I quit plenty of the too-many applications and Web pages that I always leave open.

    I also need to partition the two gleaming new Seagate 2TB drives, to follow Rick's recommended scheme:

    "I personally always use a pair of drives with 2 partitions. I boot off partition 1 on drive 1 and back it up to partition 2 of drive 2. Then I put data on partition 1 of drive 2 and back that up to partition 2 of drive 1. That gives me twice the performance for the exact same dollar."

    I had thought about making the boot partitions smaller than the data partitions, on the theory that the data may need more room than the OS and applications; but, after today's experience with the too-small partition on the SSD, I am asking your recommendations.

    I do not work with any individually large files, such as huge Photoshop files. I apparently just like keeping lots of everyday things open.

  18. #38
    Join Date
    Aug 2001
    Location
    Grangeville, ID USA
    Posts
    9,119

    Default

    Tested..... not enough time in our day to run whatever it is SoftRAID does. So we format them, set up an app that does a continuous write, read, compare cycle and walk away for a bit. If afterwards it shows zero errors, we are good to go. Amazing how often this finds an error and saves us a early drive failure, which I hate almost worse than anything.

    Firmware - I haven't a clue where the line is drawn - meaning where Apple does firmware changes making it impossible to install or run earlier OS revisions. No idea at all. Yes, you should be able to go get whatever OS installer and just plop it down on a empty partition. I guess some of that is which MacbookPro you have and what OS revisions it supports.
    molṑn labe'
    "I am a mortal enemy to arbitrary government and unlimited power. I am naturally very jealous for the rights and liberties of my country, and the least encroachment of those invaluable privileges is apt to make my blood boil."
--Ben Franklin

  19. #39
    Join Date
    Sep 2003
    Posts
    310

    Default Still Turning Off When Should Be Sleeping!

    Two brand-new hard drives an a clean install of Mac OS X El Capitan 10.11.3 later, the MacPro3,1 still turns itself off when it should go to sleep! What problems of hardware (or firmware?) could possibly cause a narcoleptic Mac?


    As before, the problem is intermittent. Sometimes the computer sleeps; others, it turns off.
    Last edited by Bozocity; 02-22-2016 at 12:27 PM.

  20. #40
    Join Date
    Sep 2003
    Posts
    310

    Default "Snap!" Upon Shut-Down

    Update, after repeated auto-shutdowns and re-boots:

    Although the problem is intermittent, most of the time, the computer shuts itself off. When the time comes for it to go to sleep (per the twelve minutes set in the Energy Saver control panel), a surprisingly loud, electrical "snap!" is heard, and it shuts off instead. It is not a controlled shut-down. One moment it's running; the next, it's powered off.

    The "snap!" is not from the computer's speakers. The computer has separately powered speakers set as the audio source. They have not been powered on (I often prefer the silence). Neither the external nor the built-in speaker is currently receiving any power or signal. The "snap!" is some other part doing what? arcing? shorting? Some switch involved in attempting sleep must be making this noise.
    Last edited by Bozocity; 02-22-2016 at 05:37 PM.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •