cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Showing results for 
Search instead for 
Did you mean: 

Community Tip - Visit the PTCooler (the community lounge) to get to know your fellow community members and check out some of Dale's Friday Humor posts! X

Linux Windchill Installation, Mirroring, Clustering and Vaults with UnionFS and F5 ARX data tiering.

patrick.chin1
1-Newbie

Linux Windchill Installation, Mirroring, Clustering and Vaults with UnionFS and F5 ARX data tiering.

We have started off testing successfully with UnionFS and applying F5 Arx to our Windchill Production System with our EVA and NetApp for data tiering. Our produciton, test and developments are all Linux RedHat Servers.


What UnionFS does is that it has the ability to instantly mirror your production system to your cluster, test and development without the need to perform another install. The benefits are:



  • no data duplication between systems (Windchill codebase, vaults, cultering, test, development)

  • System specific layered NFS mounts for service packs, customizations, system local machine specific files, cache, logs, vaults

  • instant time based mirroring by reseting the readonly layer mounts to a different time.

  • on test, development, clusteredsystems only contain a layer for modified files for server specific names and modifications of files (code development, service packs)


With UnionFS , we did the following:


Withan OOTB Windchill installation:



  • Delete All the Windchill cache and logs (tasks, logs including Apache, Tomcat, Cognos, Directory Server)

  • On the Test, Clusters, Developments

    • have the UnionFSmounts comprised main 2 NFS mounts:

      • read only of Complete Windchill Installation (production nfs export)

      • write nfs for system specific files

      • additional nfs layers for builds and service pack installations which can be easily be removed if causes issues. Thus, the only requirement is to backup the database. (REALLY COOL!) If a build or service pack causes issues, just remove the nfs layer and reset the database.



With F5 Arx, files can be sorted on various data storage appliances depending on number of times accessed versus the Windchill vaulting rules. Thus, more expensive faster storage hardware appliances can be set at tier 1 and the slower be set at tier 2 and 3.


Overall, this saves tons of money for storage and disk requirements to maintain ITIL methodology for multiple systems of production, test, development and even clustering. We have our systems on a blade infrastructure with a 10GIGE backbone switch and all apps installed on remote mounts to the storage array of EVA and NetApp. Thus, we know it requires a goodnetwork connection from the servers to the storage array which is standard these days. With the F5 Arx, the entire installation, tmp, vaults can be reused to other systems as a nfs mount. Then add the UnionFS to the mix and you have system specificfiles on a different NFS mount layer.



I need to test virtualization with this ensuring that there is at least 1 core and 4GIGs of ram per process.



Has anyone else done this? Have you been successful?

1 REPLY 1

This really benefits custoemrs with large vaults who needupdates to their systems and don't need the expense of duplicating their vaults in their test and development. The mature layered application is now AuFS.





In Reply to Patrick Chin:



We have started off testing successfully with UnionFS and applying F5 Arx to our Windchill Production System with our EVA and NetApp for data tiering. Our produciton, test and developments are all Linux RedHat Servers.


What UnionFS does is that it has the ability to instantly mirror your production system to your cluster, test and development without the need to perform another install. The benefits are:



  • no data duplication between systems (Windchill codebase, vaults, cultering, test, development)

  • System specific layered NFS mounts for service packs, customizations, system local machine specific files, cache, logs, vaults

  • instant time based mirroring by reseting the readonly layer mounts to a different time.

  • on test, development, clusteredsystems only contain a layer for modified files for server specific names and modifications of files (code development, service packs)


With UnionFS , we did the following:


Withan OOTB Windchill installation:



  • Delete All the Windchill cache and logs (tasks, logs including Apache, Tomcat, Cognos, Directory Server)

  • On the Test, Clusters, Developments

    • have the UnionFSmounts comprised main 2 NFS mounts:

      • read only of Complete Windchill Installation (production nfs export)

      • write nfs for system specific files

      • additional nfs layers for builds and service pack installations which can be easily be removed if causes issues. Thus, the only requirement is to backup the database. (REALLY COOL!) If a build or service pack causes issues, just remove the nfs layer and reset the database.



With F5 Arx, files can be sorted on various data storage appliances depending on number of times accessed versus the Windchill vaulting rules. Thus, more expensive faster storage hardware appliances can be set at tier 1 and the slower be set at tier 2 and 3.


Overall, this saves tons of money for storage and disk requirements to maintain ITIL methodology for multiple systems of production, test, development and even clustering. We have our systems on a blade infrastructure with a 10GIGE backbone switch and all apps installed on remote mounts to the storage array of EVA and NetApp. Thus, we know it requires a goodnetwork connection from the servers to the storage array which is standard these days. With the F5 Arx, the entire installation, tmp, vaults can be reused to other systems as a nfs mount. Then add the UnionFS to the mix and you have system specificfiles on a different NFS mount layer.



I need to test virtualization with this ensuring that there is at least 1 core and 4GIGs of ram per process.



Has anyone else done this? Have you been successful?


Top Tags