Mac Pro Possessed

Question

I have a Dual 2.26 Quad-Core Mac Pro. It has 6 Gigs of factory RAM and the NVIDIA GeForce GT 120 graphics card and is running OS 10.5.8. The computer’s primary use is for editing on Final Cut Pro (v 7.0.2), but it does serve day-to-day purposes as well.
Everything that can go wrong with it has and endless attempts at fixes, at best, only temporarily fix the system. There have been so many issues and so many approaches to finding the cause(s) that I’m afraid this will be a pretty long posting and I’ll still forget some things. Please bear with me. If you like a challenge, I’m your man.
I got this Mac on 10/2/09. We (Any time I write ‘we,’ I am referring to a highly-qualified Mac technician who has done much of the heavy lifting throughout this process; I think he now wants to become a farmer) copied a lot of files from my previous G5 (running OS 10.4.11), but did a new clean install of Final Cut Suite and of any other apps. We also installed 3 1-TB Samsung HE103UJ internal drives in bays 2, 3 & 4 in a RAID using SoftRAID software v.3.6.8. Gorgeous. I’ve got the best computer in the world. Ha.
On 10/6, I launch and get a SoftRAID Disk Error message telling me that one of the RAIDed partitions encountered a read error. This message will continue from time to time throughout my trials, but, until the last few days, we’ve never detected any actual problems with these drives and no tests run on them could find any performance issues.
On 10/9 “Final Cut Pro quit unexpectedly. The problem may have been caused by the KGCore plug-in.” This seems to be a known problem (http://forums.creativecow.net/thread/8/976905), which will continue to pop up from time to time, but otherwise, things are running well.
11/17, I finish one job and take over an existing Final Cut project. There are 4.5 TB of media spread over six 1-TB drives (eventually this will grow to 7 drives) daisy-chained via short firewire 800 cables to my CPU. From this point on in the narrative, my internal drives will not be used for media at all. I will continue occasionally to get read error messages from SoftRAID, but I detect no actual problem and all three of those drives are sitting idle anyway.
Lot’s of technical problems start to occur now, but many of them have to do internally within the project (I would elaborate, but I’ve looked ahead and this sucker’s going to be long enough without side issues). One issue I do address is in early January, I add an additional card to my Mac to provide more firewire 800 ports. I now have six of the external drives each plugged into their own port and only one drive piggybacking on one of the others. I see no real improvement, but still, I attribute my crashes to FCP project issues.
Then in April I start crashing and freezing. All. The. Time. Endlessly. Throughout all of this, btw, I am regularly trashing permissions, caches and temporary files; repairing permissions; trashing FCP preferences; running Disk Utility and Disk Warrior and Compressor Repair (http://www.digitalrebellion.com/compressor_repair.htm). My tech finally suggests we clear the internal RAIDed drives so we can reformat them not as a RAID and then do a fresh install of the OS on one of them and install Final Cut there in its own little isolated world. On 4/22, we do all this. That drive (just to confuse things) is called ‘Final Cut.’ The only third-party drivers running from here on out are SoftRAID’s and Kensington Mouseworks.
I sometimes have to make several attempts to get FCP to launch, but it seems fairly stable for a few days other than it’s lagging something terrible. I’ll do some extremely basic action —loading a clip into the viewer for example— and I’ll get the beach ball for 5-10 seconds. Sometimes when this occurs there appears to be no activity on the external drives (that is, the lights aren’t flashing).
4/27, This is when things start to become really inconsistent. I pull all drives and launch FCP without any media connected. It appears to run fine. No lags, no crashes. I quit and remount just two of the media drives, each into a native fw 800 port. Restart FCP. At first it lags when I’m rolling over media from one of these two drives, then after a while it quits lagging. I quit app, restart; everything runs fine. I let it sit while I write an email, return and now it lags when rolling over media from the one drive. Then it gets over it again and runs fine.
I reboot from my other startup drive and get a SoftRAID error. It is not on the Final Cut startup drive, but one I’m not actually using now. I reboot from the Final Cut drive and get the same SoftRAID error message (I may not have cleared the error messages in SoftRAID; can’t recall). I eject the two internal drives I’m not using.
I swap the firewire cables for the two drives I’ve been testing and launch FCP. Seems to run fine. Quit and attempt to remount all 7 drives. One won’t mount. This is new. Launched Disc Utility and it hung trying to find the drives. Forced Quit while it was still looking. Tried to restart the Mac, but it hung up on shutdown. Forced shutdown from power button, powered down drives.
Powered up. Launched FCP. Ran fine for a while then encountered occasional lags. Sometimes the problem drive from earlier flickers during these hang-ups and sometimes none of the drives do.([I realize, in the history of computing, ‘occasional hang ups’ would not be considered an issue, but on the original boot drive, the app either crashed or it didn’t; when it ran, it ran immediately.)
Had to go back to the other startup drive for something. Came back, relaunched FCP. Had a small sequence (2:44) in timeline, hit play, it hung up. First one drive flickered then several other drives in turn. Some of them had no media in this sequence. Eventually it settled down. Everything feels very unstable.
Ran Disk Utility and Disk Warrior. The iffy drive’s directory was only slightly off, less than 5%.
4/28, While reading an email, with nothing going on in the background, I got a SoftRAID read error. That drive holds only files for apps that weren’t running at the time. Since these drives are no longer RAIDed, I can isolate which drive is having problems. It’s ‘Media 1’ in Bay 2. I quit FCP and it crashes on shutdown. I’ve now got a great app called FCS Maintenance Pack (http://www.digitalrebellion.com/fcs_maintenance.htm), which includes a Crash Analyzer. It suggests this crash was caused by Qmaster.
4/30, Since the main issue now seems to be hang time. We decide to install a couple of eSATA cards. I plug four of the seven drives into eSATA ports and now have just 3 drives running each to their own firewire 800. Possibly a marginal improvement, but probably not.
5/3 We run some speed tests and the external drive that was acting up earlier is pretty slow. Turns out it was very fragmented. We defrag and things look better. Much better, in fact. For an entire day. It runs so well that I decide, on 5/4, to run Compressor on a ninety minute show to see if I’ve got an any performance improvements (I decided to not an entire Compressor/Qmaster conversation here). Compression was much faster. Then, not immediately, but shortly thereafter, I had a panic crash where the screen comes up in several languages telling you to shut down from the power button. I restarted the computer to make sure it would start up and then quit for the day.
5/6-7. Today, it was very difficult to get the computer to start. As best I can recall, it would stop halfway into the startup screen and I would have to power down at the CPU and start over. Eventually it got up and after trashing some preferences seemed to run OK. Today and tomorrow everything works perfectly. No drive issues, no lag, no SoftRAID error messages. My computer works exactly the way I would expect the biggest, baddest Mac on the market should. I take the weekend off.
Monday 5/10/10, Early in the day, I get this error message: “The device you removed was not properly put away. Data might have been lost or damaged. Before you unplug your device, you must first select its icon in the Finder and choose Eject from the File menu.”
I haven’t pulled anything, nor do I see any drives missing from the desktop. Something could have gone offline for an instant and then back on without me spotting it, but I can’t find any evidence. This has happened a few times over the past week or so (forgot to include above and I can’t say exactly when it started). At least once when this message came up it locked up Final Cut. Still, everything runs well today. Tomorrow I’ll only make modest use of the computer and not run FCP at all.
Wednesday, 5/12/10, FCP starts to hang up again. I run speed tests on the drives and one of them — not the one I recently defragged — just hangs during what should be a quick test. I eject and remount the drive and it gives me decent — not great — performance numbers. I go back to work. The rest of the day is rough with lots of lagging. Then, in the early evening, shortly after I’ve run Disk Warrior and Disk Utility, I’m playing a sequence and I hit the pause button and it doesn’t pause. Typically, in FCP, if it crashes, it crashes, everything stops. Not this time. I think my keyboard has become disconnected. I crawl under my desk and fiddle with the USB connector and when I come back up, my screens (I have two) were all messy, like acne, like snow, like nuclear rain. Scattered pixels over the app’s windows and over the desktop were screwy.
Eventually FCP crashed (Crash Analyzer just said it was caused by an unknown error) and I was able to regain control of the computer. I would like to quit for the evening and start drinking, but I’ve got to get out a DVD of my project. I eventually successfully export a QuickTime Reference file from Final Cut and then successfully compress it in Compressor. I then eject all the external drives. The QuickTime is on “Media 1,” which is one of the three internal 1-TB Samsung drives.
I start the build in DVD Studio Pro. It gets all the way to the end, hangs on 'Completed,' and then I get the "…device you removed was not properly put away…" message. Now I can't swear that one of the drive icons didn't blink off for a moment, but they were all there when I looked. But when I opened Media 1 all the folders in it appeared to be empty.
A restart brings all that info back. I check Console for the moment of the incident and, sure enough, it indicates “disk4: media is not present.” I think this is referring to “Media 1,” but I can’t find in Disk Utility or SoftRAID a reference to disk4 so I’m not sure.
5/13/10, I’m crashing all day long. Crash Analyzer, variously says the problems are caused by my graphics card, a corrupt project file, some inappropriate media in the project, corrupt audio files or a problematic QuickTime component. There were also some force-quits, which don’t generate a report. Now, I’m pretty impressed with this analytic software overall, but I think I’ve overwhelmed it. I don’t honestly think it’s a different one of these things each time I crash.
I run Apple Hardware Diagnostics from the install CD and it finds nothing wrong.
5/14/10, Today I start a crash journal. The remainder of this history is pasted from it.
I wish I had started this when the problems started. We now join our story already in progress.
5/14/10, 2:45 FC Drive
Pulled “Media 1,” which appeared to be the source of the SoftRAID errors, from computer then relaunched and started FCP. Played OK for a few minutes then started displaying pink or green frames. Then it hung for a while. I did other things for a few minutes. Returned and now it played, but with HUGE pixilation and loud audio static. Then hung up. I'm now hoping it will crash — which will generate a Crash Report in Crash Analyzer — instead of having to force quit — which won't generate a report.
3:05
Forced quit CompressorTranscoder and CompressorController from the Activity Monitor. No affect. Forced quit all Compressor and Qmaster related processes. No effect. Finally forced quit FCP.
3:09
Launched Disk Utility and as it displayed drives, I got the "…the device you removed was not properly put away" message. All drives appear on the desktop, but “Deception” (one of the external eSATA-connected drives) doesn't show in Disk Utility. Here are a few lines from the Console for that moment (bolding mine):
May 14 15:08:54 martin-nelsons-mac-pro Disk Utility[364]: Disk Utility started.
May 14 15:09:00 martin-nelsons-mac-pro kernel[0]: Failed to issue COM RESET successfully after 3 attempts. Failing...
May 14 15:09:00 martin-nelsons-mac-pro kernel[0]: IOBlockStorageDriver[IOBlockStorageDriver]; executeRequest: request failed to start!
May 14 15:09:00 martin-nelsons-mac-pro kernel[0]: IOBlockStorageDriver[IOdisBlk8s3:ock Stresourceor agshoeDriverrt];age ex.
May 14 15:09:00 martin-nelsons-mac-pro kernel[0]: ecuteRequest: request failed to start!
May 14 15:09:00 martin-nelsons-mac-pro mds[80]: (/Volumes/ Deception /.Spotlight-V100/Store-V1/Stores/FEE1E744-878D-4219-9CEF-A037A6F8F98D)(Error) IndexCI in ci_ftruncate:ftruncate(53 /Volumes/Deception/.Spotlight-V100/Store-V1/Stores/FEE1E744-878D-4219-9CEF-A037 A6F8F98D/live.2.indexPostings, 4096) error:22
And then farther down:
May 14 15:09:00 martin-nelsons-mac-pro kernel[0]: disk8s3: media is not present.
May 14 15:09:00 martin-nelsons-mac-pro kernel[0]:
May 14 15:09:00: --- last message repeated 1 time ---
Etc.
When I clicked on “Deception” in a Finder window, Finder quit responding. Quit all open apps and tried to do a force relaunch of Finder. Desktop cleared and then everything hung there. Chose Shut Down from the Apple pulldown menu. Menu went away leaving only the desktop image with no icons on it, but it didn't shut down. Forced shutdown from the CPU 3:31
Powered off all external drives. Restarted. Accidentally in other startup disc. Restart again on FC drive. Can't find any reference in SoftRAID, Disk Utility or System Profiler to disk8s3 (from Console text above). Mounted Deception and still couldn't find a reference. 3:59
4:03, Decided to verify “Deception” using Disk Utility. It's predicting it will take 28 minutes. Should take about 20 seconds. I verified all drives including this one yesterday and found no problems. Also ran Disk Warrior and found it less than 10% (maybe much less; don't remember) out of order.
Meanwhile, Disk Utility seems to have stopped progressing. Activity Monitor doesn't suggest it's hanging up, but there's no movement in Real or Virtual Memory usage and in Disk Utility nothing happening in the progress bar. I stopped verify and started it again. It makes a quick run to Checking volume bitmap and then again says Estimated time: 28 minutes. That's just crazy. 4:14, I'm gonna ignore it for a while and see if there's any progress. 4:22, Finished running finally. “Deception appears to be OK.”
Launch Disk Warrior. Deception is 8% out of order. Try to rebuild, but am told it's in use. Quit all apps including DW. Restart DW. Deception is sill supposedly in use. Restart computer. This takes forever. Launch DW and attempt to rebuild Deception's directory. I get a disk malfunction error and am told to backup my data from the preview disk
I've got a spare drive. Guess I'll start copying.
4:54 Started copying 931 GB
This ends up taking a couple of days, which gives me time to write this novel.
5/16, After copying. My tech guy wants me to run TechTools v3.1.3. It hangs up trying to find volumes. I have to force quit. I then try to shut down, but ultimately have to do it from the CPU. I power off all external drives and restart. It won’t restart. I power down from the CPU and start again. This time after 10 minutes or so it starts. I think it might be a good idea to restart. It shuts down, but won’t restart. I launch from my other startup disk. This works fine. I run Disk Utility and Disk Warrior on the “Final Cut” startup disk. It finds no problems and only minor directory issues. I rebuild the directory anyway.
I restart in “Final Cut” drive and rebuild its permissions. Launch TechTools and it scans volumes quickly. It’s now running. It’s now done. It finds nothing.
More general notes:
The system log has hundreds of pages of this error:
May 13 16:55:23 martin-nelsons-mac-pro [0x0-0x105105].com.apple.FinalCutPro[1676]: Final Cut Pro(1676,0xa02aa720) malloc: * mmap(size=268439552) failed (error code=12)
May 13 16:55:23 martin-nelsons-mac-pro [0x0-0x105105].com.apple.FinalCutPro[1676]: * error: can't allocate region
May 13 16:55:23 martin-nelsons-mac-pro [0x0-0x105105].com.apple.FinalCutPro[1676]: * set a breakpoint in mallocerrorbreak to debug
May 13 16:55:23 martin-nelsons-mac-pro [0x0-0x105105].com.apple.FinalCutPro[1676]: Final Cut Pro(1676,0xa02aa720) malloc: * mmap(size=268439552) failed (error code=12)
May 13 16:55:23 martin-nelsons-mac-pro [0x0-0x105105].com.apple.FinalCutPro[1676]: * error: can't allocate region
There were major crashes at various points when all of the firewire drives were offline and other times when they were online and two of the three extra internal drives were offline. The “Final Cut” drive, which is one of the internal, aftermarket 1TB drives and is serving as the startup drive when working in Final Cut Pro, was, of course, online.
There have been numerous times in the last few weeks when I have not been able to successfully turn on the computer on the first try —it gets partway launched then hangs, forcing me to hit the power button— or turn off the computer without resorting to hitting the power button on the CPU.
Possible fixes:
We’ve removed the one drive flagged by Softraid for errors, Media 1. And I’m replacing the second drive that appears damaged, Deception. Can the problem be further disk errors or driver conflicts? Most of my external drives are Lacie 1TB disks, formatted with the Lacie utility in Mac OS Extended format with the Apple Partition Map. The two remaining internal Samsung drives are formatted with SoftRAID as single disks, in Mac OS Extended format with GUID partition scheme and block sizes optimized for digital video. Memory was tested with “memtest.”
Could it be bad RAM? We have the stock 6GB that came with the system. Could replacing it with 16GB of new RAM possibly fix it?
Is it a bad video card, processor, or motherboard problem? The pervasiveness of the problem makes me think it is a hardware issue, not software.
I’d like to test the drives with a second Mac Pro, but I have none available.
Would downgrading Final Cut Studio and removing one of the Pro Applications Updates possibly help?
Well, I'm exhausted. I wonder if anyone made it to the end. If you have any ideas I would be eternally grateful. I'm tapped.
Martin

Answer 1

Most likely culprit is your Lacie drives which are notorious for flaky power supplies which cause just the sort of erratic behaviour that you describe. Have had similar issues myself. Lacie has been good about replacing my power converter-- twice. If you have seven of them, good luck finding the bad one or more. Like I said, the behaviour is erratic. Sometimes it'll work sometimes not. I can't comment on Samsung drives because I've never used them and have never heard anything about them. Same with Softraid. Why not just use Disk Utility. Sounds like you're ready for an external hardware raid. This will be much faster, much more reliable, and may even be cheaper than seven firewire drives. Check out G Tech, and Maxx Digital for starters.
Best of luck
Terry