ESXi

 View Only
  • 1.  Bad blocks on a datastore - a fix?

    Posted Jun 24, 2015 03:14 PM

    We have a VMware 5.0.x host on a Dell PowerEdge 610. Our monitoring systems says that the RAID array has problems - the disks have bad blocks (Puncturing bad block). Our VMs are experiencing performance problems and we are not able to copy the vdmk files due to read problems.

    As advised from Dell, we ran the following commands from the /opt/lsi/MegaCLI directory:

    ./MegaCli –FwTermLog –Dsply –aALL > /tmp/lsi.txt

    ./MegaCli -PDList -aALL > /tmp/disks.txt

    ./MegaCli -LDInfo -LALL -aALL > /tmp/LDinfo.txt

    ./MegaCli -AdpAllInfo -aALL > tmp/Adapterinfo.txt

    ./MegaCli -AdpBbuCmd -aALL > /tmp/Battery.txt

    See results attached. Dell advised us to replace all disks in the array - that means for us recreating the datastore and restoring our VMs. We could do this from previous backups, but before we do this would like to know if there is a chance of repairing the current datastore? Is there such a thing as a checkdisk for vmware, so that the bad blocks are marked etc. so that we can then get a complete copy of our vdmks - which we would then transfer to a newly created datastore?

    Thanks for any help on this urgent issue.

    Oliver



  • 2.  RE: Bad blocks on a datastore - a fix?

    Posted Jun 24, 2015 05:29 PM

    I am not sure about removing the bad block from vmfs5 but this article is really nice , I would recommend you to read

    VMFS Extents - Are they bad, or simply misunderstood? | VMware vSphere Blog - VMware Blogs