Closed dannyk96 closed 6 months ago
I have reverted to 5.1.4.0 (was 5.1.5.0) to see if this was the cause.
I have found that it works the first time it is run, but fails if re-run.
(I re-ran as /usr/local/bin was missing from root's PATH so it couldn't run ansible-playbook
Somehow there is state left over from the previous install, even though I did an rm -rf /usr/lpp/mmfs
I beleive the cause is that the rpm gpfs.bda-integration-1.0.7-0.noarch
gets installed.
a re-run of ./spectrumscale will always fails as it goes down the wrong rabbithole if it detects this RPM !
workaround is to remove gpfs.bda-integration-1.0.7-0.noarch
Proper fix is for the developers of ./spectrumscale to fix their code!
Daniel, I'm not sure I understand what you exactly did and what "it works the first time it is run" means.
Especially difficult as I'm not seeing such an issue with my tests.
Please note that using the spectrumscale
command within the deployed VM is not supported. Seems you tried that for troubleshooting / to complete a stopped deployment, correct?
Can you please describe a sequence like:
vagrant up
vagrant ssh
into the VMcd /usr/lpp/mmfs/5.1.5.0/ansible-toolkit
and executed ./spectrumscale ...
?
Closing due to missing feedback / unsupported operation. Please re-open or create a new issue with the missing information if there should still be an issue.
The installer
/vagrant/install/setup*.sh
is generating many error mesages of the formThis appears to stop ./spectrumscale from setting any parameters, nor to run 'install' Maybe this is because it is Centos not Redhat ?