NetBackup for VMware error 13
Hi All,
I post this previously:
https://www.veritas.com/community/forums/netbackup-vmware-error-13
It took quite long time to solve even with the Symantec support, then I am sorry I forgot to post the solution earlier. Actually it has nothing to do with the network, I don't really undersand the detail, but I understand the idea: t has something to do with the LVM disk configruation.
Basicaclly, the problematic RHEL had improper configuration of the LVM. The system was partitioned but we use the whole partition instead. This one is actually improper configuration of LVM disk, according to Veritas this may cause bpbkar / bpbkar32 to crash while running VMware Backups.
WARNING!!! this may cause corruption on your disk, you better backup the system first, I don't guarantee it will work on you, better consult with Veritas first.
The solution by Veritas:
Running xxd on the disk device with the suspected invalid configuration will help confirm the issue.
Example:
Command: xxd -l 100000 /dev/sdc
The output when the LVM is improperly configured will appear similar to:
00001e0: 0000 0000 0000 0000 0000 0000 0000 0000 ................
00001f0: 0000 0000 0000 0000 0000 0000 0000 55aa ..............U.
0000200: 4c41 4245 4c4f 4e45 0100 0000 0000 0000 LABELONE........
0000210: e595 f5aa 2000 0000 4c56 4d32 2030 3031 .... ...LVM2 001
0000220: 4730 4276 466e 3854 4c6d 7577 3767 6264 G0BvFn8TLmuw7gbd
0000230: 706a 5034 6641 6a43 5931 3331 3274 6535 pjP4fAjCY1312te5
RedHat (RHEL) states; if the disk is LVM(LVM disk) then its first sector is filled with zero's . However, in this case the disk has LVM signature as well as the MBR signature and one partition (sdc1) - which is invalid as per LVM specification:
Caused:
https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/6/html-single/Logical_Volume_...
4.2.1.1. Setting the Partition Type
If you are using a whole disk device for your physical volume, the disk must have no partition table. For DOS disk partitions, the partition id should be set to 0x8e using the fdisk or cfdisk command or an equivalent. For whole disk devices only the partition table must be erased, which will effectively destroy all data on that disk. You can remove an existing partition table by zeroing the first sector with the following command:
# dd if=/dev/zero of=PhysicalVolume bs=512 count=1
Here we can see that both the MBR signature (55aa hexadecimal in the first sector and LVM signature (LABELONE) are on the second sector of the disk. This configuration, while not supported on RHEL, seems to cause no problems with the OS, however NetBackup sees this as inconsistency.
Because of this configuration LVM+MBR, instead of taking /dev/sdc1 it takes /dev/sdc which is why our offset gets calculated incorrectly.
NOTE THIS PROCEDURE ONLY APPLIES TO /dev/sdc DISK's THAT ARE IMPROPERLY CONFIGURED, IT IS NECESSARY TO CHANGE THE COMMAND TO REFLECT THE CORRECT DISK.
To fix this inconsistency it is necessary to zero down the 2nd sector. Please clone the VM or take snapshot to which VM can be restored and keep a copy of the same just in case something goes wrong.
Solution:
Please run the following commands:
# dd if=/dev/sdc of=/tmp/dump.img bs=512 count=1
# dd if=/dev/zero of=/dev/sdc bs=512 count=2
# dd if=/tmp/dump.img of=/dev/sdc bs=512 count=1
# sync
On my case it works.
Regards,
Iwan
