Both the internal and external drive are about a month old, bought at the same time. HGST Deskstar 4TB 7200 RPM SATA III 6Gbps, 64MB Cache. External drive is in a FW800 enclosure.
So what could possibly have gone wrong here?
I guess I could pull the machine apart again and swap the two drives. I'm not sure what else to try.
It seems very unlikely to me that one of these drives is faulty when the other is not, especially as it doesn't seem to exhibit any other failures (nothing in logs, I/O seems fast, etc.)
But even if it's a motherboard problem that just happened to manifest when I swapped drives, it is still an I/O problem so why can't I prove that with numbers?
It wouldn't be surprising for one of two drives to fail early. These guys have experience with lots and lots of hard drives and have kept stats on their experiences.
http://blog.backblaze.com/2013/11/12/how-long-do-disk-drives-last/
That said, when you boot off the backup drive, is it on a different bus? Seems to be - external is firewire, internal is sata. Bad sata cable? Or maybe it's getting too hot inside the computer?
Backing up what Kevin says, I had a Macbook Pro where the internal HDD failed three times in a year (each time it was losing data). Turns out it was the SATA cable (which is a pain to replace on the Mid-2009 Macbook Pros). We worked this out by using Scannerz (scs-online.com) on the drive connected to the internal SATA and then again via USB. From reading comments since then, it seems that this is an emerging problem on Macbook Pros...
Well, it's the same internal cable that was in there before. I'm reasonably sure I seated it properly, though, because this is not my first rodeo. (If only it didn't take over an hour to disassemble and reassemble iMacs...)
This was the page that helped me... Maybe it has some info which could help you.
https://discussions.apple.com/thread/5172609?start=0&tstart=0
SATA cables aren't spec'd for a huge number of removal/reinsertion cycles, and the cables can develop intermittent problems from other causes -- exceeding the bend radius while fiddling with them, etc. They go "bad" sooner and more often than you might expect.
Swapping the cable is a pain, but probably slightly less painful than swapping the drive. 10:1 a new cable fixes it.
Charles
SATA uses CRC to detect when packets didn't make it across the wire OK. SATA drives will report these CRC errors as one of the SMART statistics (the one named "UDMA_CRC_Error_Count" -- if you're wondering why "UDMA", it's because the feature was first introduced on UltraDMA parallel ATA, and they never changed the name).
jwz's drive is reporting zero CRC errors. It's probably not the cable.
Unless I'm very much mistaken, the drive's SMART stats will only report packets received at the drive with failed CRCs. It won't capture packets that were corrupted arriving at the SATA controller -- and since those go over a different pair of conductors, it's entirely possible the cable is intermittently flaky in only one of the two directions. The resulting required retransmits would explain the stuttering.
It may or may not be the cable -- but the cable is definitely the first thing I would try swapping.
Do you happen to know what kind of cable is required for a mid-2010 iMac11,3 with a 4TB SATA3 drive? I assume SATA cables are not all the same, and I'm having a hard time googling up that answer.
Actually, all "normal" SATA cables for 2.5" and 3.5" drives (i.e. not SATA-e, or SFF8087-to-SATA breakout cables, or other oddities you won't find inside your Mac) are the same; the only technical differences are whether they have a locking clip or not, and whether they have straight or right-angle connectors. I've not seen the inside of your Mac, but get the same end configuration as you currently have. All cables made in the last few years should be fine for SATA3 (6Gbps) use, but I'd avoid the absolute cheapest yum-cha cables.
If you have a local computer hardware parts place you like, whatever they have in their house brand is likely to be okay. If you'd prefer to mailorder, any of Monoprice's cables are fine as long as they meet your length requirements. Don't get one that's a foot longer than you need. These, depending on length and quantity, are in the $2-3 range.
http://www.monoprice.com/Search?cp_id=10226&cs_id=1022602&keyword=sata3
Well that was fun.
I swapped the backup drive into the machine: still skips! So it's not the drive.
I was gonna replace the cable, but the other end disappeared under a board that I couldn't figure out how to remove.
Hooray.
The problem not following the drive is good info; after eliminating the drive, it is much more likely to be the cable than the SATA controller port on the mainboard. It could also be a driver problem but I don't know offhand if the Intel MB SATA drivers changed in your OS point-release upgrade.
Regarding getting at that cable... I don't actually know what "Imac 11,3" as a model number means, so I don't know the exact machine you're referring to.
Would you be so lucky that it's one of the Imac models listed in the teardown/repair guides at Ifixit?
http://www.ifixit.com/Device/iMac_Intel
Re: the bidirectionality of the SMART statistics, I don't have the relevant standards in front of me so I can't say definitively. But I'm fairly sure SATA retries bad packets, which requires some sort of ACK/NAK scheme, which means that the drive should know when packets it sends aren't received correctly.
Got smartmontools installed? If so, what's smartctl -A /dev/disk0 say?
It says... stuff...
ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
1 Raw_Read_Error_Rate 0x000b 100 100 016 Pre-fail Always - 0
2 Throughput_Performance 0x0005 136 136 054 Pre-fail Offline - 80
3 Spin_Up_Time 0x0007 168 168 024 Pre-fail Always - 485 (Average 436)
4 Start_Stop_Count 0x0012 100 100 000 Old_age Always - 18
5 Reallocated_Sector_Ct 0x0033 100 100 005 Pre-fail Always - 0
7 Seek_Error_Rate 0x000b 100 100 067 Pre-fail Always - 0
8 Seek_Time_Performance 0x0005 100 100 020 Pre-fail Offline - 0
9 Power_On_Hours 0x0012 100 100 000 Old_age Always - 940
10 Spin_Retry_Count 0x0013 100 100 060 Pre-fail Always - 0
12 Power_Cycle_Count 0x0032 100 100 000 Old_age Always - 8
192 Power-Off_Retract_Count 0x0032 100 100 000 Old_age Always - 54
193 Load_Cycle_Count 0x0012 100 100 000 Old_age Always - 54
194 Temperature_Celsius 0x0002 115 115 000 Old_age Always - 52 (Min/Max 21/54)
196 Reallocated_Event_Count 0x0032 100 100 000 Old_age Always - 0
197 Current_Pending_Sector 0x0022 100 100 000 Old_age Always - 0
198 Offline_Uncorrectable 0x0008 100 100 000 Old_age Offline - 0
199 UDMA_CRC_Error_Count 0x000a 200 200 000 Old_age Always - 0
As far as I can tell, most of that llooks fine, but the temperature at 52 degrees (ID# 194) is at the upper end of what's safe for a harddrive (although the manufacturer specs say 60).
I'm afraid the next course of action is probably to try the backup drive plugged-in internally so you can narrow it down to either the disk or the interface.
According to the RAW_VALUE colum you have zero reallocated sectors (ID #5 Reallocated_Sector_Ct, #196 Reallocated_Event_Count), and there are zero sectors in the middle of going bad or being spared out (if there were you'd expect #197 or #198 to be greater than 0). Additionally, zero CRC errors (#199) means that you probably do not have a bad cable.
So there's nothing from the SMART report suggesting the usual reasons why drives might time out for a while, which would've explained your video stuttering. On the other hand at least it looks like the disk is healthy...
It's not CoolSpin branded is it? I've seen a buttload of complaints about those spinning down every 30 - 90 seconds, then having to spin up again.
All this st-stuttering b-business seems to be holding up a much-ancitipated mixtape. So, here are some bands I've enjoyed:
Prides, "Out of the Blue":
Zella Day, "Sweet Ophelia":
Jungle, "Busy Earnin'":
The Twoks ("twahks", not "tooks"), "240 Volts":
The Twoks, "240 Volts":
Forgot the best one. Stunning performance by Jena Irene Asciutto of her song "Unbreakable Me":
> Unfortunately, a bunch of things have changed on this machine over the last few months, and I'm not sure exactly when this problem began: I've upgraded from 10.9.1 to 10.9.2; replaced the RAM; replaced the internal drive. So there are several things that could have been the cause. It could be some new stupidity in the Quicktime library, or a disk performance problem, or... well, maybe it can only be those two.
This sounds very familiar - I had those symptoms AND I had display corruption on my iMac that I could only reproduce occasionally. Apple store tested it for 24 hours (their claim) and found nothing wrong. Got tired and sold that piece of crap. So much for it just works.
I think you might want to try putting in a SSD in place of that drive - given most of Apple's hardware now ships with SSD - may be they stopped caring about the rotating drives.
If there was such a thing as a 4TB SSD, I surely would, but there is not. The whole point of this upgrade was that my 2TB drive was full.
i looked at both posts and didn't notice you confirming the drives have identical firmware. also noticed that HGST takes a user-hostile stance regarding firmware unlike WD and Seagate. anyway, that's the first place i would have looked after verifying throughput on each port of the disk controller.
http://www.hgst.com/support/faqs >> "Is it possible to obtain a firmware upgrade for my hard drive?"
Since I can't get a firmware update, I suppose the fact that I have no way to find out what my firmware version is is academic. Their tool that tells you only runs on Windows.
Possibly viewable with native system profiler:
a) http://www.zdnet.com/yes-you-can-update-your-macs-hard-drive-firmware-in-os-x-7000008328/
b) http://support.seagate.com/firmware/firmware_update_procedure_mac.html
If they differ, then there's your ammunition for contacting HGST support. Even if they're the same, it's worth writing and saying you're experiencing nonsensical throughput issues and asking if there's new firmware.
And as always, kill hit---err, back up before applying.
PS - WinDFT is not distributed as a portable, but it is. I installed and dragged the binary to a windows liveCD image; booted said image from a flash drive; and WinDFT ran without incident. So: totally possible to run it without installing windows.
FYI: the reason i suspect firmware is because your problems sound like cache or NCQ bugs/failures. if you have hdparm in OS X, you might be able to see some metrics playing with -Tt, -Q, and --direct.
check out this guy's woes: https://www.linuxquestions.org/questions/linux-hardware-18/how-to-test-hdd-internal-cache-4175493396/#post5111702
> a 4TB SSD
Ok, but you can always test with a smaller SSD. If the videos did not stutter on the SSD, that would prove my point that the issue has something to do with OS X's handling of rotating drives.
Well, here ya go. http://www.sandisk.com/enterprise/sas-ssd/optimus-max-ssd/. No guarantees it will solve your stuttering problem, of course. That's due to gremlins.
Perhaps when booting from the external drive, the OS begins using the external driving for logging, swap, etc., reducing I/O contention. Does the stuttering always stop when the video is on a different spindle from the booted partition?
I meant: both booted off of and playing videos off of the external drive. Internal drive not mounted.
You know, an idea for trouble shooting which you may very well hate would be to swap out another drive and install another OS, like Windows or Linux JUST FOR TESTING, then try to stress the IO on both the internal port and the external port. If all goes perfectly smoothly, then MacOS is not configuring or using the internal SATA port right suddenly, and you at least have something to look at. If it gives the same or similar issue, you know it's the mobo. I know I know, you hate MS and Linux is for the BDSM crowd, but I only suggest it for testing, so, like, a couple hours tops.
You're adorable.