"Error reading Pool Data Block ########, status=20652" when using NSS volumes

  • 3793433
  • 17-Apr-2007
  • 27-Apr-2012

Environment

Novell Open Enterprise Server (Linux based)

Situation

Pools and volumes will not mount.
Using PowerPath for multipathing
NSSMU under devices will show multiple paths to the same pool
In /var/log/messages you may see following errors:
"Error reading Pool Data Block ########, status=20652"
"Error reading Pool Data Block ########, status=20206"

Attempting to mount volume you may get a -1 error, unable to mount volume.
In nsscon you may see errors:
"Unable to correct super block header 3"
"Unable to correct super block header 4"


Another customer reported the following in /var/log/messages repeatedly when trying to mount a pool.
  kernel: attempt to access beyond end of device
  kernel: dm‑13: rw=0, want=314560528, limit=104855520 
  kernel:  BIO_UPTODATE not set.  Got an IO error
  kernel:  BIO_EOF set.  Got an IO error

Resolution

NSS is trying to mount one of the paths to the Pool, it was not mounting the emcpower* path that is needed to bring this up.
Fix is to edit the /etc/evms.conf file and make the following changes. In the include make sure you include anything you want evms to see, ie if your boot device is an evms device you will need to add it here as well (ie. include = [ emcpower* sda ] ).
sysfs_devices {
include [ emcpower* ]
exclude = [ ]
}

Additional Information

After correcting this and rebooting server, if the pool had been expanded across multiple disks you may need to run a pool rebuild in order for it to return to an active state.

Feedback service temporarily unavailable. For content questions or problems, please contact Support.