A bolt from the blue
By Rajeshr-Oracle on Oct 08, 2006
Landed in Mumbai airport at around 11:00 A.M. this morning. The flight journey was made interesting by a sailor sitting next to me (sorry that I am unable to recollect his name). In just under half an hour time he briefed me on his experiences travelling all around the world and how miserable life could be in a ship with only 21 sailors managing the whole ship. It seems, at times, they don't see a glimpse of land for 3 months at stretch. Quite a scary thought I believe.
I checked in at the hotel at around 11:45 A.M or so and reached here (training venue) at 12:30 sharp. I came here to set up the lab for tomorrows class. I have done this several times before. And I was (over)confident to complete my job in a couple of hours time. I expected a text book installation, but didn't have a clue that a surprise was in store for me and I were to have a very tiring and tension filled afternoon. Let me explain:
Generally I carry a Flash Archive in my laptop to the training venues so that I could install all the lab bundles for the course on the student machines. And the flash archive for IDM 345 course contained Solaris 9 Operating Environment, mysql installed and the lab bundles for the course. And a clean installation (read extraction) of this archive ensures a smooth take off for this course , since all the student machines would be in a 'ready to go' mode. And I was confident that my Flash Installation would not take more than two hours time. I had ten Sun Blade 1500 machines at my disposal for performing this installation and I chose one of them to be my jumpstart server. Let me also mention here that the archive that I had in my laptop was created on an Ultra 10 machine. I have had arguments with my training participants on several occasions on archive installation on machines with varying architecture. And until this afternoon I was under the strong impression that an archive created on one architecture could be installed on any architecture without any problem. It didn't take much time for me to create a jumpstart server and I was using a Solaris 10 CD as the boot CD on the jumpstart server so that the clients could boot from the CD on the jumpstart server. The installation procedure appeared faster than ever and I was busy thinking on my agenda for the evening once I reach my hotel after what appeared to me as a clean installation process. It was a joy to watch nine machines extracting my the flash archive from my jumpstart server. Once the installation got over, one machine rebooted and then it crashed with an error indicating that the root filesystem is corrupted and needs fsck. I felt slightly shocked, but had the confidence that I would be able to repair the problem. But then I saw even the rest of the machines crashing one by one. It was a depressing sight to see each of those machine going into the maintenance mode one after the other. Perhaps it was an early Diwali dhamaka. I immediately called up Samson Selwyn, a very senior colleague of mine on his hand phone. Well, this is certainly not the first time that I spoiled his weekends. I have done that before and done it on several occasions. So I think he's used to it now. He gave me a few tips to somehow get out of the bad situation that I was in. I was feeling very tired and hungry and hence went out to gobble something. Back in my mind, this was really bothering me, for IDM 345 is a lab intensive course and any minor upset in the lab setup would spoil my next five days.
I came back and tried booting the clients using CD and very soon discovered that the system was behaving very abnormal. The format command gave me "No disks found" error. I found myself at sea and the only solution that I could think of was to create an archive of Solaris 10 OS of my Jumpstart Server (Sun Blade 1500) and use the same to install the OS 'at least' on all the client machines. By this time, I had already reached a conclusion (whether right or wrong, I don't know) that an archive created on one architecture may not work on other architecture. Well, at least this time the installation of the OS went through fine. So now all the machines in here are installed with Solaris 10 OS. Now I will have to request my participants to carefully install the other softwares tomorrow while doing the lab. That will tax a lot of our time, but then I don't seem to have a choice. And I really hope that things work fine tomorrow. It has been a very tiring and exhaustive day and all I feel now is to somehow get back to my hotel and take some rest.