Hi @ClouderaAdmin ,
Thanks for reaching out to Cloudera community. You should stop the cluster for doing the IP changes. See below for some recommendations which may help you:
============
Below document makes the following assumptions:
- Hostnames remain unchanged
- Nodes are not added or removed from the cluster
- OS or other software remain the same
- The entire cluster is shut down while the change is performed
The process should be the same for a secure cluster, as long as name resolution works fine.
Changes to be made across the cluster:
- Alter all hard-coded IP addresses.
- If /etc/hosts is used for name lookups, ensure that all hosts in the cluster are updated.
- If DNS is used for name lookups, allow for TTL expiry so that the new IP addresses are returned to queries.
Zookeeper:
- Nothing specific to worry about if host name lookups are working correctly and quorum membership only has hostnames
HDFS:
- Replica availability will be unaffected as the DNs will report in the new IP address when sending heartbeats
- Blockpool name might contain the older IP address but it's just a label and can be left alone
- Exclude and allow host lists should be updated, refer to dfs.hosts and dfs.hosts.exclude for file locations
- If racks are configured, update the topology.data file
Cloudera Manager:
- Host to role mapping will remain valid.
- The agent will send in the node's new IP address in the heartbeat.
- Ensure that /etc/cloudera-scm-agent/config.ini on each node contains the correct hostname of Cloudera Manager node (preferred to a hard coded IP address).
Note: All other services not mentioned above should continue functioning as is.
Hope this helps.
Li
Li Wang, Technical Solution Manager
Was your question answered? Make sure to mark the answer as the accepted solution.If you find a reply useful, say thanks by clicking on the thumbs up button.
Learn more about the Cloudera Community:Terms of Service
Community Guidelines
How to use the forum