fishbacon

Q: I need help troubleshooting a Mac Pro 5,1.

Hi everyone, I manage the IT and network operations at my company and I've run into a particularly troublesome Mac Pro and could use some help. Here's some background on the machine and problem:

 

  • Mac Pro 5,1
  • 6 core 3.33GHz
  • 32GB RAM (Samsung)
  • ATi Radeon HD 5770
  • Latest build of Yosemite

 

I used this machine with no problem for a few months just doing heavy file transfers from external drive to network storage. The problem is that after transferring it to the office where it's used for video editing the machine suffers from random crashes while being used where it doesn't freeze or anything but instead just instantly powers off and restarts. The programs used that seem to cause the crash are Adobe Premiere CC 2014/2015 and Final Cut 7.

 

Here's what I've done so far that has made no difference: I've run memtest on the RAM for a week straight until the full test was completed and it passed with no errors. I've tried other RAM that I know to be good. I installed a new hard drive. I removed all extra PCI cards like RAID, USB 3, etc. I load tested it for 3 days at 99% CPU and it never crashed or had a problem. All that was left in the machine was the ATi Graphics Card. I reinstalled OSX. I moved the machine to the side and turned off all power saving features and it stayed running for 16 days continuous with no restarts. We tried to use it for work again and within 2 hours the crash returned.

 

I'm a bit lost as to what is going on here, I built an identical machine before this one and it's had no problems so far and I've run out of troubleshooting ideas. Does anyone have any suggestions as to what I should try next? Thanks.

Mac Pro, OS X Yosemite (10.10.5), 32GB RAM

Posted on Apr 22, 2016 7:58 AM

Close

Q: I need help troubleshooting a Mac Pro 5,1.

  • All replies
  • Helpful answers

  • by lllaass,

    lllaass lllaass Apr 22, 2016 8:09 AM in response to fishbacon
    Level 10 (188,143 points)
    Desktops
    Apr 22, 2016 8:09 AM in response to fishbacon

    If you are getting a kernel panic log it and post it here:

    Mac OS X: How to log a kernel panic - Apple Support

     

    Otherwise:

    • Note the time of the next unintended reboot. After rebooting open the Console app in Finder>Applications>Utilities and look for log entries at the time of the reboot. Post suspicious ones here.

    Look before the log entry that starts with BOOT_TIME since that is the first line of a bootup

  • by Grant Bennet-Alder,

    Grant Bennet-Alder Grant Bennet-Alder Apr 22, 2016 8:44 AM in response to fishbacon
    Level 9 (60,774 points)
    Desktops
    Apr 22, 2016 8:44 AM in response to fishbacon

    Mac Pro with Xeon processor uses Error Correcting Code memory. Memory problems do not fester undetected in these Macs. Correctable errors are corrected on-the-fly in one stretched memory cycle, and uncorrectable errors, such as double-bit errors cause a distinctive kernel panic, machine check.

  • by fishbacon,

    fishbacon fishbacon Apr 22, 2016 9:44 AM in response to lllaass
    Level 1 (8 points)
    Desktops
    Apr 22, 2016 9:44 AM in response to lllaass

    I'm not seeing a panic log there but I was able to pull the latest crash log if that helps, I can't seem to paste it without losing the formatting but here's a downloadable/viewable version:

     

     

    https://drive.google.com/file/d/0B_O6SF3r-SFZSzZwUU5Fak8yaUU/view?usp=sharing

  • by Grant Bennet-Alder,

    Grant Bennet-Alder Grant Bennet-Alder Apr 22, 2016 9:56 AM in response to Grant Bennet-Alder
    Level 9 (60,774 points)
    Desktops
    Apr 22, 2016 9:56 AM in response to Grant Bennet-Alder

    That is not a kernel panic, that is an Application panic. Its report is only interesting to the developer.

     

    If a kernel panic occurs, information is added to a log file in the folder /Library/Logs/DiagnosticReports

     

    .

  • by fishbacon,

    fishbacon fishbacon Apr 22, 2016 10:26 AM in response to Grant Bennet-Alder
    Level 1 (8 points)
    Desktops
    Apr 22, 2016 10:26 AM in response to Grant Bennet-Alder

    Is it possibly not a kernal panic as I am not seeing any logs for that?

  • by John Lockwood,

    John Lockwood John Lockwood Apr 22, 2016 10:28 AM in response to fishbacon
    Level 6 (9,265 points)
    Servers Enterprise
    Apr 22, 2016 10:28 AM in response to fishbacon

    As the problems started after the machine was physically moved it could be that has caused the problem. While it is less common on a Mac Pro I would check the power lead at both ends to make sure that it is firmly seated and undamaged. As an example the old style Mac mini had a power connector that was prone to being pulled out by the weight of the power adapter.

     

    Similarly due to being moved internal components could have become partially unseated. I would therefore when it is turned off open the case and make sure the RAM chips are pushed in firmly, the video card is also checked and pushed in to its slot and any internal cables especially the video power cables and fan cables that you can get at are also all checked.

     

    If the problem only happens when you are heavily pushing the CPU usage e.g. when doing a render then overheating could be an issue if a fan has failed or is as above not connected properly. For programs that heavily push the use of the GPU chip i.e. video card then if the video card does not get enough power it will shutdown. A video card draws up to 75w of power from the PCIe slot but might also need additional power from one or two PCIe power cables. The Mac Pro supports a maximum of two 6-pin power cables each of which can also provide 75w giving a maximum total therefore of 215w. If one is not connected correctly to the video card or the video card needs even more than 215w then this could cause a problem.

     

    It would also be worth running Disk Utility and checking the SMART status of the hard drive(s).

  • by fishbacon,

    fishbacon fishbacon Apr 22, 2016 11:00 AM in response to John Lockwood
    Level 1 (8 points)
    Desktops
    Apr 22, 2016 11:00 AM in response to John Lockwood

    Thanks for the suggestions, I should add those to my original posting since I've also tried all of those also. The GFX card only uses one cable and even tried swapping in another. Fans are all good, I installed SMC Fan control to monitor temps and everything has been good on that end even when I load tested it.

  • by Grant Bennet-Alder,

    Grant Bennet-Alder Grant Bennet-Alder Apr 22, 2016 11:09 AM in response to fishbacon
    Level 9 (60,774 points)
    Desktops
    Apr 22, 2016 11:09 AM in response to fishbacon

    Is the northbridge chip temperature "crazy-high"?

     

    This can be caused by failure of the retaining pins. It is simple to check for on a one socket (4 or 6 -core). After powering off, slide out the processor shelf and try to wiggle the chip in the center of the board. The northbridge peeks out from under the adjacent processor heatsink. It should seem solid. If it is obviously wiggly, you have problem.

  • by fishbacon,

    fishbacon fishbacon Apr 22, 2016 1:58 PM in response to Grant Bennet-Alder
    Level 1 (8 points)
    Desktops
    Apr 22, 2016 1:58 PM in response to Grant Bennet-Alder

    I checked it out and it seems solid. I'm using istat and I don't see a temp listed for Northbridge just the power usage (1.11V). The temp for the CPU A is 160 degrees Fahrenheit and the Heatsink for that one is at 89, do those seem reasonable for a machine not being used and just running?