Forum Discussion
NimbostratusMigration from i5800 to r5800
NacreousHello,
My approach would be to a create new tenant in r5800 assign the same vlans as current cluster but with new self ips.
Add these new self ips to whatever fw rules you may have for the old ones.
Then add the new tenant as member of the existing cluster.
Sync and confirm that new node is able to communicate with as many pools as the active one.
During a maintenance window failover to new node and confirm functionality.
If everything is fine proceed with the other r5800
At the end you will have a cluster of 4 and you can start decommissioning i5800.
Another option is to shutdown standby i5800, use it's self ips to new r5800 tenant and add it to the cluster. Confirm again connectivity, failover, check functionality etc.
This way you do need new ips.
Nimbostratus
Hello injeyan,
Thanks for your reply. I hope your doing well.
I was expecting some step by step procedure for migration but in the reply I see only for self ip migration anyhow thanks for your reply.Is there any chance to get migration document and details about rseries 5800 model how differ from i5800 wrt gui interface and other parameters tab.Also, any guide to transfer each partition migration.
I also studied one of F5 document that rseries doesn't have HA between hardware box instead it has tenants redundancy with HA is this correct like what are the differences between i5800 and r5800.
Please help me if any clear document for migration i5800 to r5800 for reference.
Thanks in advance.
Regards,
Moula Ali
- Injeyan_Kostas
Hi Ayish
In the scenario I described, you only see the Self IPs because you’re essentially creating a new, empty BIG-IP with fresh Self IPs, adding it to the existing cluster, and then syncing it. This means there’s no need to manually migrate any configuration, it will all be synchronized automatically.
Regarding the differences between iSeries and rSeries:
- rSeries runs F5OS, which is a completely different operating system that acts as a hypervisor.
- On top of F5OS, you deploy BIG-IP instances (called tenants).
- Like any hypervisor, you must first configure all networking in F5OS (interfaces, trunks, VLANs, etc.) and then assign the VLANs to the BIG-IP tenants.
F5OS itself is standalone, so as you correctly noted, there is no HA at the F5OS layer. However, HA is fully supported at the tenant (BIG-IP) level, just as with traditional BIG-IP deployments.
If you’ve worked with vCMP before, the concept is almost identical.
Regarding partitions, these will also be synchronized as part of the cluster sync process.
For the initial configuration of F5OS, you can refer to:
https://techdocs.f5.com/en-us/hardware/f5-rseries-systems-getting-started/gs-system-initial-config.htmlOnce the BIG-IP tenant is created, the workflow is exactly the same as you’re already familiar with.
