Cluster incorrectly upgraded with non-cluster install.sh

[Rich Graves]

...how can I recover?

One node of a single-server one+one RedHat cluster (RHEL4U6 64-bit) was upgraded from 4.5.5NE to 4.5.6NE using zcs/install.sh, without following the documented cluster upgrade procedure. Among the problems this caused was slapd not starting and the conf symlink to mountpoints being moved out of the way and recreated as a directory.

We are currently running with just one node, and getting errors in the admin GUI (presumably because of failure to communicate with the second node). How can we upgrade or recreate the second node?

FWIW, error when trying to browse Server Status at server:7071 is:

Message: system failure: java.lang.StringIndexOutOfBoundsException: String index out of range: -11
Error code: service.FAILURE
Method: ZmCsfeCommand.invoke
Details:soap:Receiver

Btw, install.sh should be modified to detect clustering and abort, thereby preventing this user error.

[jholder]

Hi Rich,
I've been following your support ticket, hence I didn't reply.

As far as the installer, see bug:
Bug 18426 - installer should detect cluster install