Big Data Appliance X4-2 Hardware
 

PLA-Support>Eng Systems>BDA>Big Data Appliance>DB: BDA_EST
 

In this Document

Goal

Solution

References

Created from <SR 3-14057197351>

Applies to:

Goal

Steps to recover/restore a corrupt/bad USB drive when it is detected during reimage/reboot of a BDA server.

Some background information can be found in: BDA Boot Process Frequently Asked Questions (FAQ) (Doc ID 2042579.1).

Solution

There are a few options for recovering/restoring a corrupt/bad USB drive.

1. Shutdown the server and re-seat internal USB drive. This may resolve the issue. If not continue to 2.

2. Try the steps in: Rebooting a BDA Server Leads to a Black Screen with Blinking Cursor Instead of the Grub Menu (Doc ID 2142573.1).  It may be possible to workaround the issue by booting from the hard disk first.

If you can boot into the hard disk by switching the BIOS boot order to boot, then use the utility bdaredousb to reprovision the USB.

If this does not help, continue to 3

3. Try to recover/restore the corrupt/bad USB drive using a "good" external USB drive.

The idea here is to copy the image from an internal USB on a "healthy" server onto an external USB and then copy the image from the external USB back to the internal USB on the affected node.

You can find the internal USB using lsscsi. For example from the lsscsi output below the internal USB is /dev/sdm: 

You can find the external USB (once it is plugged into a server) using lsscsi or with "showusb".  For example from the lsscsi output below after plugging the external USB drive into a "healthy" server the external USB is /dev/sdn: 

or try running the script showusb which tries to detect an external USB and shows what device it is on.

The detailed steps are below.  Perform them as 'root' user on the affected node unless specified otherwise. 

a) Plug an external USB drive into one of the "healthy" BDA servers. Basically use any server but the server with the corrupt/bad USB being recovered.

b) Use dd to copy the image from the internal USB drive drive on the "healthy" server to the external USB drive.

The general syntax is:

For example based on the above lsscsi output:

Output is like:

c) Plug the external USB drive into the affected node.

d) Boot the affected node into rescue mode using the Linux iso image using JavaRconsole.

See: Steps For Re-imaging a Single BDA Server Using CD-ROM Redirection (Doc ID 1559505.1). Go to the bottom steps around step 13. Or see: Steps to Boot Oracle Big Data Appliance Node into USB Rescue Mode using Grub Menu (Doc ID 1570447.1).

e) Once the node is up restore the internal USB drive by copying the USB image from the external USB drive to the internal USB drive.

The general syntax is:

For example based on the above lsscsi output:

Output like: 

f) Perform another reboot to boot into the internal USB drive.

g) If the failure was detected during reimage, go on to reimage the node following: Steps For Re-imaging a Single BDA Server Using CD-ROM Redirection (Doc ID 1559505.1) if it is not possible to boot into the OS. If you can boot into the OS follow: Oracle Big Data Appliance Base Image Version 4.5.0 for New Installations or Reprovisioning Existing Installations on Oracle Linux 6 (Doc ID 2173434.1).  Or if you want to reimage to the original factor ship network settings, see How to Reimage a BDA Server in Rescue Mode to Original Factory Ship Network Settings for a 4.5 and Lower Base Images (Doc ID 2227956.1).

Since the reimaging uses the internal USB to store the installer (and that is now a regular rescue image) it will be necessary to use single-server reimaging to reimage this node.

This solution has no attachment