June 30, 20179 yr I have a single cache drive in Unraid. HPA is enabled. I would like to disable the HPA. What would be the recommended method. I tried: hdparm -N /dev/sdi Command line comes back and tells me HPA is enabled. Any ideas? Could I simply remove it from being selected as cache drive, preclear the drive, and re-select it as the cache drive? Thanks
June 30, 20179 yr The most convenient method is to simply use the SetMax feature of HDAT2 https://www.hdat2.com/ HDAT2 is mentioned in the link above that outlines the issue, but it says you need a CD drive to use it -- that is no longer the case. Just download an ISO from the HDAT2 site; then download Rufus, and you can create a bootable flash drive that will let you easily run HDAT2. https://rufus.akeo.ie/
June 30, 20179 yr Leave the HPA partition on the drive. It barely drops the capacity of the drive, and more importantly anecdotal evidence shows that if your server's BIOS put an HPA partitition onto the drive, then it will NOT put it onto any other drive. HPA is only an issue if it winds up on the parity drive(s)
June 30, 20179 yr Author Too late Remove it. Looked in syslog found the current and native sizes. Jun 30 12:23:37 unraid kernel: ata8.00: HPA detected: current 468860015, native 468862128 Then used hdparm to remove the HPA. hdparm -N p468862128 /dev/sdi hdparm -N /dev/sdi Removed the HPA. Though I munged something else as the GUI shows no dockers, and my VM's listed in the GUI appear to be a set I was running a year ago. Going to search the forum to see about finding a similar symptom.
June 30, 20179 yr You did copy everything off of the drive before you ran the command, right? If not, your best bet at this point would be to set the HPA back exactly the way it was. Removing the HPA probably caused the drive to be unmountable.
August 9, 20178 yr Author Wasn't sure how to properly migration. Was under the assumption that a cache drive was truly a cache like one would find on a CPU or similar as buffer with high speed flash or the like. That assumption was wrong. The result by removing in the method I did, was a server config with hostnames and the like of over a year old. Was able to recover by removing the docker, VM, configurations, adding "new" VM definitions in the GUI using the *,img disks and re-applying dockers as new. Edited August 9, 20178 yr by joedotmac
Archived
This topic is now archived and is closed to further replies.