Just curious if anyone else has seen this before.. I tried to use NCM to roll out a IOS upgrade to several Cisco 2960-S switch Stacks. I chose to use the following command archive download-sw /leave-old-sw tftp://${StorageAddress}/' + @IOS_FILENAME. The file I am using is a .tar file because I don't know the size of each individual stack and the command line above will get the new file to each switch in the stack. The /leave-old-sw was a requirement from the boss otherwise I would have used the /overwrite.
I also had the last command as reload in 30 minutes.
Everything went fine with the test group of switch stacks but after the reload the switch stack was still running the old code. The log files in NCM showed the download took place and the command output log showed that each switch in the stack received the download. When I checked the switch stack and did a dir /all I noticed that my new image file was in a subdirectory flash(#):/update/ . Each switch in the stack had the same subdirectory.
Is this behavior normal? Do I need to change my script to have a 'no system boot:' and then a 'system boot: flash1/update/new_ios_file' , and then the reload command?
If I change the boot Variable will the startup for the first time on the new image create the directory structure I am use to seeing Like
flash1:/ios_Image/ios_image.bin
My preference would be to use the archive command and the .tar file because I am trying to get away from downloading the .bin file to each individual switch stack member...