Add special scripts for Data Manager in Puppet and have them deployed in the cmsadmin nodes
Move special scripts for Data Manager in Puppet and have them deployed in the dCache Tape Chimera node.
After this is done Dave Mason should test that it works and we should move pool 'cmsstor111' back to the production branch.
This should be assigned to Chih-Hao (but this redmine project is not yet fully setup).
#2 Updated by Gerard Bernabeu Altayo over 4 years ago
We discussed this a bit more during the meeting: Chih-Hao will create an RPM with the scripts, any password will go in a config file via puppet eyaml and this will be deployed on cmsadmin nodes (will need to talk to Tim and add D. Mason in cmsadmin as a user).
#6 Updated by Chih-Hao Huang over 4 years ago
- Tracker changed from Bug to Feature
- Due date changed from 06/10/2015 to 07/29/2015
- Status changed from New to Assigned
- Start date changed from 06/08/2015 to 07/22/2015
- % Done changed from 0 to 20
 Dave Mason is already in root's .k5login every where, there is no need for special permission treatment.
-- it has to be one of us as root to run these scripts
 cmsstor111 is already on production branch for a while -- no special branch needed any more
 git@cms-git:cms-dataops is created for keeping track of these scripts (they used to be in cvs)
-- they are under cms-dataops/file-family-creation
-- no more concern of "what if cmsstor111 dies suddenly"
 they could be deployed through git since they could be placed any where on a tape store node
-- they only require (1) enstore in command path and (2) /pnfs mounted root writable
 will create rpm though -- even though, it can not be deployed to any node due the requirement in 
 recording the process in redmind
[*] file family creation is an occasional task.
#8 Updated by Chih-Hao Huang about 4 years ago
- Status changed from Assigned to Resolved
The entire working directory, /root/FF, including the original scripts, Dave's scripts and historical log files, has been migrated to cmsstor261.
The migration was done through tar so the time stamps were kept.
Dave M. has been using it for two weeks.
#10 Updated by Gerard Bernabeu Altayo about 4 years ago
- Status changed from Resolved to Assigned
- % Done changed from 100 to 10
thanks for moving the scripts to a node that won't get decommissioned in the short term but bear in mind that this is just a mid term patch. You still have to get this scripts in a well known repository (eg GIT) and deploy them in a managed way (eg: RPM or puppet).
If you put them in an RPM following the standard DCSO practices you will see that the files become GIT managed 'by default'.
I'm reopening the ticket.