8 Replies Latest reply on Nov 4, 2008 9:37 AM by Bill Robinson

    VPC setup Best Pratice

      Hi

       

      I am trying to setup VPC and seems to work okay on one app server but when it goes to the other app server it always fails. I was talking to support and mentioned when you have multi app server environment the VPC gets installed on the first server and then somehow shared to the other app servers in order for this to work, If that is the case what is the Best Practice for this. Can I ut on a NAS share or setup that it only runs on that first server and if so how.

       

      The repository with patches exists on the NAS share.

       

      Thanks, Ken

        • 1. Re: VPC setup Best Pratice
          Bill Robinson

          there's not really a best practice doc or anything - and the vpc as it is not is not designed to work in a multi-appserver environment.

           

          to make it work, you need to mount the /usr/nsh/patch directory on all your appservers (nfs, etc)

           

          if you're using AIX patching, you need to mount the bff_root directory across all your appservers (whatever path you set that to)

           

          if this is on windows, it may be more problematic because you can't setup persistant mounts on windows like you can w/ unix.

          • 2. Re: VPC setup Best Pratice

            yeah, for windows multi-appserver environment, you might need to setup some scheduled task or something to sync the patch directory.

            I am in the same boat at my client right now. Maybe this is a good point to bring up during arch discussions if the client is pushing for a all windows environment and is also thinkin about vpc.

             

            Does anyone have the future roadmap for patching?

            • 3. Re: VPC setup Best Pratice

              I should have mentioned it is a RHEL4 app servers....i figure out the nfs mount situation and try that

              • 4. Re: VPC setup Best Pratice
                Bill Robinson

                if you can do the nfs, that's the easiest i think. what we've been trying to do is get a large space allocated on SAN for the 'file server' and then mount that under like /opt/bladelogic for example. then we create 2 folders under there /opt/bladelogic/storage and /opt/bladelogic/vpc, then use the storage dir for the normal file server and share out the vpc dir over nfs and put the /usr/nsh/patch dir under there and also the various patch repositories.

                • 5. Re: VPC setup Best Pratice

                  I actually put the share on the bldepot since bith app servers can see it, I was able to run analysis jobs from both servers without an issue then a couple more runs and stopped working...I keep getting syntax error while parsing conf file.

                   

                  Nothing has changed and even recreated the conf file....redhatrepo.conf

                   

                  ANyone seen this before....

                  • 6. Re: VPC setup Best Pratice
                    Bill Robinson

                    which conf file? there's a few, can you post logs from the job run that show the error ?

                    • 7. Re: VPC setup Best Pratice

                      Not sure....it doesn't specify in the log so I thought the redhatrepo.conf

                       

                      Here is the contents of the log...

                       

                      Error Nov 4, 2008 10:04:38 AM Syntax error while parsing conf file

                      Info Nov 4, 2008 10:04:34 AM Started running the job 'RHEL4ES-32-U5 RedHat Patch Analysis Job' on application server 'lctsp170'(2,008,000)

                      Info Nov 4, 2008 10:04:35 AM Number of Servers to process in parallel per platform: 10

                      Info Nov 4, 2008 10:04:35 AM Host List File: *528943*

                      Info Nov 4, 2008 10:04:35 AM Prefix = 528943-lctsp170

                      Info Nov 4, 2008 10:04:35 AM hostlist is /usr/nsh/tmp/scripts/job_528943/hosts528945

                      Info Nov 4, 2008 10:04:35 AM nexec lctvp083 'sh -c "test -f /etc/redhat-release"'

                      Info Nov 4, 2008 10:04:35 AM nexec lctvd108 'sh -c "test -f /etc/redhat-release"'

                      Info Nov 4, 2008 10:04:38 AM Options:

                      Info Nov 4, 2008 10:04:38 AM Debug Flag: ON

                      Info Nov 4, 2008 10:04:38 AM Policy Name: RedHat Patch Analysis.

                      Info Nov 4, 2008 10:04:38 AM Hosts Specified: YES

                      Info Nov 4, 2008 10:04:38 AM Working Directory: /bldepot/vpc/patch/rhpu/Scripts

                      Info Nov 4, 2008 10:04:38 AM Mode: Analysis Required

                      Info Nov 4, 2008 10:04:38 AM Mode: BLPackaging Required

                      Info Nov 4, 2008 10:04:38 AM Repository Tags: my_repo

                      Info Nov 4, 2008 10:04:38 AM Analysis type: repo, Details:updates

                      Info Nov 4, 2008 10:04:38 AM Repository Tags: my_repo

                      Info Nov 4, 2008 10:04:38 AM Depot Folder Name: /Patch Analysis Items/RedHat Patch Analysis/Patches

                      Info Nov 4, 2008 10:04:38 AM Job Folder Name: /Patch Analysis Jobs/RedHat Patch Analysis/Patch Deploy Jobs

                      Info Nov 4, 2008 10:04:38 AM Could not open /bldepot/vpc/patch/rhpu/Scripts/jobrpmexcludelist.txt, Skipping Considering Exclude File!

                      Info Nov 4, 2008 10:04:38 AM Change dir /bldepot/vpc/patch/rhpu/Scripts

                      Info Nov 4, 2008 10:04:38 AM Exit Code 1

                      Info Nov 4, 2008 10:04:38 AM The job 'RHEL4ES-32-U5 RedHat Patch Analysis Job' has failed

                      • 8. Re: VPC setup Best Pratice
                        Bill Robinson

                        is that w/ the debug on?

                         

                        you might have to dig through the perl script and see where it's generating that message from - grep for that message in the rhpc.pl file ?