Client install issues
Hope someone can point me in the right direction on this..
We have a number of servers that we are migrating from an outside managed hosting provider, over to being managed by us. The systems in question are running Solaris 10 / Sparc. After migrating, it was noticed that the clients were running netbackup 6.5, so I am attempting to upgrade them to what we are running in our environment (7.1), not only to get them current but to also overcome the zfs bug in running pre-6.5.2 NBU.
The problem is, it seems that when they have "installed" the netbackup client on the systems, all they did is take a tarball and explode it out on the servers. When i attempt to install the client (using install_client_file sftp <host> <user>) and attempt to run the sh /tmp/bp.xxx/client_config it runs thru the installer (does PBX first), but bombs out because it doesnt find the SYMCnbclt package on the system.
I was able to get it up to 7.0 (it was the only version of the client install i had at the time) by running the full NB7.0 client install script, but it still had some errors on the install(didnt properly config vnetd service) and in doing a pkginfo | grep -i netbackup, it still does not know of any netbackup being installed. Although it does temporarily get me around the zfs issue (all_local_disk) and i can back up the system, i am not able to install any newer clients until a solution can be had.
So, it seems that the issue is that the installer sees that netbackup is installed on the system, but since it doesnt live in the Solaris package db, it fails to install
Below is what i have tried so far:
moved /usr/openv/netbackup to another location and attempted to install
moved /usr/openv to another location and attempted to install
Does anyone have any ideas how to overcome this issue?
I found the solution and it was so stupid, it is almost embarassing to give it.. But after running the install a couple more times using the full 7.1 Client install tar.gz and looking line by line at the trace file, i came across a line that said "./cp_to_client: pkgadd: not found". Seeing this, it sent me towards looking to see why pkgadd was not found. I checked my pathing and sure enought, /usr/sbin/ was not pathed on this system as it is for all the systems that we had managed... After fixing the pathing and rerunning the install, it worked properly. I have not yet tested the install from the "install_client_files" push yet, but i don't doubt that will work as well...
I would think that the script should check to see if pkgadd exists in the path and if it doesnt display an error message that states this.. The error messages that if had displayed were pretty vague in troubleshooting..