Hi Guru's I did not receive any responses to my query, but I managed to solve the problem in any case, with some help from the vendor for the Disk subsystem. I also logged a P1 Call with Sun, and though this equipment are not covered by our support contract, they gave me some assistance with advice. Unfortunately, they claimed that this is simply NOT POSSIBLE. Yet, I managed to recover the disk fully. The solution was to do the following: 1. Use the HDS "Shadow-Image" function to create a "backup" of the published LUN - just in case I screw it up. 2. Create a Completely new OPEN-E*x-SUN LUN having the same geometry as the original, and make it available to the host. 3. Discover the new LUN on the host (reconfigure reboot required in most cases) 4. Use "format" to label both the original and the new LUN. The shadow backup LUN is not touched. 5. Use vxdiskadm to initialize the new LUN. 6. compare the prtvtoc output for the original and new LUNs At this point I have to note that part of what saved me was the fact that all my disk encapsulation has been using defaults - no "special" tweaking. If I had used special tweaking and had not documented it, it might have been hard or impossible to re-create the VTOC. 7. Pipe the prtvtoc from the "new"-ly initialized disk onto fmthard to write a new VTOC to the original disk. 8. vxdctl enable 9. vxdisk list (voila!) In the end, the HDS shadow image was just a backup - in case I screwed the LUNs up. However it worked without any problems. Still, I say rather safe than sorry. One thing I did not mention in my previous email is that our NetBackup master was one of the affected systems. Our recovery team focused on two parallel objectives: Recover the master and/or recover the disk LUNs to avoid necessitating a restore. We had just started importing tapes manually into NetBackup (using newly allocated disk) when I managed to successfully recover the LUNs. Regards, _Johan Johan Hartzenberg/G To: sunmanagers@sunmanagers.org IS/CSC cc: @CSC Subject: Lost Disk Sent by: sunmanagers-a dmin 16/07/2002 10:31 AM Hi, I have a severe problem and hope someone out there can help me. It seems like one of the NT sysadmins by accident wrote signatures to the SAN disk allocated to some UNIX systems. My first question is how can I confirm this. Secondly, is there a way I could try to copy the Solaris signature from one disk on one of the systems which are still fine, to one of these "corrupted disks" in the hope that the content of the LUN has not yet been erased? The LUNs were under Veritas volume manager control. The systems are a mix of Solaris 2.6 and Solaris 8. The disk subsystem is HDS 9960 and the LUNs are formatted as HITACHI-OPEN-E*5-0117 (the "5" varies depending on size of the expanded LUN) Some of the LUNs used to be protected by SANtinel, which lost it's configuration shortly prior to this incident - we still need to work out why. _Johan _______________________________________________ sunmanagers mailing list sunmanagers@sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagers _______________________________________________ sunmanagers mailing list sunmanagers@sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagersReceived on Fri Jul 19 03:19:58 2002
This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:42:49 EST