Disk don't always spin down

  • I just found this thread were the problem with jbd2 is also described. As you said, it seems to be an inherent problem from EXT4. I think that I'll move to XFS. Have to order a new hard drive for my main computer, so I'll use that one to back up the data, reformat the server HDD, and then give it a shot. In any case, if RAM is an issue I can probably upgrade to 4 GB (although that will be a tight squeeze in my thin-ITX server).


    I'll try over the weekend and report back my findings. Thanks for the help!

    Custom mini-ITX build
    Coolcube Mini, Intel Desktop Board DQ77KB, Intel Core i7-3770S, 8 GB DDR3 Ram, 64 GB Trascend mSata SSD (OS), X3 1TB HDD pooled + parity

    Dell Optiplex 960 sff (deprecated) - link


    Dell Optiplex FX160 (repurposed) - link


    "If you can't find it in Google, it simply doesn't exist!" - The Internetz


  • Journaling just means that there was access on the disk. You need to enable Block Dump Reporting with:
    echo 1 > /proc/sys/vm/block_dump


    And then look in /var/log/syslog for what is accessing your drive. It will gives you lines like this:
    Jan 15 21:48:29 openmediavault kernel: [136750.727828] Plex Media Serv(16049): dirtied inode 19005559 (CloudAccess.dat.tmp.XXKRFvj3) on sdc1


    Which show exactly what accessed the disk. In this case it was Plex accessing the CloudAccess.dat.tmp... file on sdc1.


    Or you can use the automated shell script I wrote which does this for you:
    My Guide to Debugging Disk Spin-ups


    Randomly disabling services might not get you where you want to go. For example with SMART set to standby and it won't wake disks to test them, but you still have a smart log in /var/log/smart.log which updates every 30 minutes with what smart is doing, even if it didn't wake up disks to test them. If your smart log is on the disk in question you'll still have smart waking it up every 30 minutes to update its logs.


    Also, be aware that turning on block dump reporting will cause a lot of information to be written to the /var/log/syslog file. I have that directory mounted in memory (tmpfs) with the flash memory plugin. But if it's on a physical disk and the one you are trying to get to spin down, it will actually interfere with you getting the disk to spin down.

  • It seems that Plex was the one keeping the HDD awake. I disabled the service 2 days ago, and now the HDD is spinning down as expexted. I'll move Plex to the RAM also and see how it works.

    Custom mini-ITX build
    Coolcube Mini, Intel Desktop Board DQ77KB, Intel Core i7-3770S, 8 GB DDR3 Ram, 64 GB Trascend mSata SSD (OS), X3 1TB HDD pooled + parity

    Dell Optiplex 960 sff (deprecated) - link


    Dell Optiplex FX160 (repurposed) - link


    "If you can't find it in Google, it simply doesn't exist!" - The Internetz


Jetzt mitmachen!

Sie haben noch kein Benutzerkonto auf unserer Seite? Registrieren Sie sich kostenlos und nehmen Sie an unserer Community teil!