Thanks that helped me.
A strange thing is that uninstall/install on primary node works fine.
But once I tried to install on secondary, I needed to clean registry
The registry part helped me, and there was another thing, the pnservice were installed in a wrong cluster group, I changed that before I moved control to secondary node, I should not do that.
When I waited to move it until it was installed on secondary node everything worked fine
I now have one issue left with this installaion.
When I move the cluster group from primary node to secondary node, then it tries to start all services.
But it fails to start the 2 cells, mcell_admin and mcell_pncell_xxxx
It fails with service-specific error code 97
What does that indicate ?
When I reboot the primary node, everything starts fine on secondary and vice versa
But it should be possible to move it without having to reboot
In Cluster machine, PNService will take care of starting/stopping server along with moving from one node to another.
Also use pw commands to start/stop process instead of services.msc window. It is highly recommended that not to use services.msc window to start/stop any of the BPPM processes.
Thanks & Regards,
That is correct, I only try to start the services because pnservice don't
When I manually move the cluster group to secondary node, pnservice fails to start the 2 cell services.
If I reboot the primary node, the move is done ok and all services are started.
It is only when I manually move cluster control to secondary node it fails, and only for those 2 services.
Afte a while it timeout and start everything on the primary node again.
In case of windows cluster installation, you could face two issues.
First you have to attach the shared disk where PNet is installed (KA349916), then you have to manually create few services (mcell_admin, mcell_pnet, mcxa, mcxela, ibrsd). Please follow the resolution below, then it rocks !
Knowledge Article ID:
During a fresh installation of BPPM on a secondary node, the DLL file msvcp71.dll does not get created. As a result, the services mcell, event adapter, event log adapter, and IBRSD are not created in the secondary node.
1 Copy the DLL file msvcp71.dll located at C:\Windows\SysWOW64 on the primary node to the same location on the secondary node.
2 Create registry entries for the missing services mcell_admin, mcell_pncell, mcwela, mcxa, and ibrsd in the secondary node. Use the following steps to create registry
A In the primary node, open Windows Registry Editor by executing the command
B Select HKEY_LOCAL_MACHINE -> SYSTEM -> Services.
C Search for the missing services. On each of the missing services, right-click and
select Export, and save it to your local drive in .reg format.
D Copy the saved services to a folder in the secondary node.
E From the secondary node, double-click each of the copied services. This updates
the secondary node’s registry with the new services.
3 Restart the secondary node.
We recently did an installation of BPPM 8.6.02 in a Veritas clustered environment and experienced the exact same problem. What we did during installation:
- Run the installation on primary node 1
- Integrate Oracle DB
- Do not integrate UDDI
- Do not integrate Blade Logic
- Do not integrate CMDB
- Install Package Repository
When doing this after the installation is complete, the PnService is Failed and BPPM is not running. What we did not do is restart the BPPM application from CLI (pw system start), ensure the PnService is Online, stop BPPM (pw system stop), ensure the PnService then went Offline, THEN move the Quorum drive (cluster group "Cluster Group" /move:ToNode002).
I am VERY certain that because of the one thing, it lead to MULTIPLE other issues, mainly the upgrade issue we experienced as well as you.
When trying to confirm this "correction", I ensured that node 1 was the owner, started BPPM from CLI (pw system start), watched the PnService come Online, ran the command to stop BPPM, watched the PnService go Offline and then moved the Quorum drive to Node 2 via the command above. Once this was done without any cleansing of the registry, I moved the shared storage to node 2 and ran the following command from CLI:
setup.cmd -DCUSTOMDB=oracle -DINSTALL_TYPE=SECONDARY
This alleviated the upgrade message we received and allowed us to complete the installation.
My advice is to install to PRIMARY, ensure that after installation on the primary node is complete the PnService is Offline, not Failed, move the Quorum to node 2, move the shared storage to Node 2, run the SECONDARY install. While I cannot guarantee success, I believe this will prevent issues with what you and I have experienced.
Hope this helps.
We struggled with clustering setup for some time. Got it installed and working, performed testing & eventually decided not to use clustering with the current release. For us, the decision boiled down to these two articles:
BMC ProactiveNet Article (Pnet must be in cluster group)
2. Can the Pnet application be configured in a separate cluster group (not cluster group)?
This is not supported in 7.7. This would be a Request for Enhancement.
MS Clustering Article (Don’t put stuff in cluster group)
The cluster group contains resources that are related to cluster administration only and are considered core resources. If you add additional resources to the group, you may not be able to administer the cluster effectively. If resources in the cluster group fail, failover may be triggered, or you may not be able to administer the cluster effectively. Therefore, Microsoft recommends that the quourm resource remain in a group that has core resources only. For best results and high availability of the cluster, do not add resources to the cluster group. Instead, put additional resources in other groups as appropriate.
Did you ever get this link from support?
We implemented clustering of BPPM in 8.6.02 using the instructions provided in the documentation, but the customer requested that we migrate the installation to the instance they created specifically for BPPM. Of course, the installer does not recognize the instance and would not install the application under this instance, so we had to manually migrate the installation to the instance and make changes to the installation, based on the documentation above. We were able to do this successfully and have BPPM installed and configured for the instance, per Microsoft's best practice, as noted above in your links.
For those needing instruction, the link provided does contain some information on installing BPPM in a clustered environment, but the base install instructions are provided in the 8.6.02 documentation, Getting Started Guide. Those instructions should be followed and if following Microsoft best practice for clustering, use the link above.