Installation

If any issues related to HPE Ezmeral Data Fabric as primary storage for Unified Analytics occur, inspect the HPE Ezmeral Data Fabric addon and review the bootstrap pod logs to identify the cause of the issue.

• To inspect the HPE Ezmeral Data Fabric addon (EzAddonDeployment), run:

  kubectl describe -n ezaddon-system ezaddondeployment storage-edf

• To review the bootstrap pod logs, run:

  kubectl logs -n ezaddon-system deploy/hpecp-bootstrap-storage-edf

  The following table lists log message examples and the reason for each message:
  TIP

  You can resolve all issues in the following table by following the steps in Preparing HPE Ezmeral Data Fabric to be Primary Storage for HPE Ezmeral Unified Analytics Software.

  Cause	Log Message Example
  The ticket provided has an expiration date.	Parsing given MapR ticket...​ The provided ticket expires at <timestamp>, please provide a ticket that does not expire​.
  The ticket provided is not a tenant ticket. For example, it may be a service ticket.	Checking if the MapR ticket is a tenant ticket...
  The CLDB hosts are not reachable. For example, due to a firewall. ​	Using maprlogin to authenticate with username/password..​. Unable to connect to any of the cluster's CLDBs. CLDBs tried: <list-of-cldbs>. Please check your cluster configuration.​
  MAPR_EXTERNAL is not configured properly and HPE Ezmeral Data Fabric reports internal IPs that are not accessible​.	Checking if mount prefix /ezua exists..​. 2024-07-18 10:49:44,0671 :5926 Timing out request 2345.234 waiting to xmit ​binding ips are: <ip-address>:7223 ​ 2024-07-18 10:49:44,0673 ERROR Cidcache fc/cidcache.cc:5333 Thread: 2337 Received error Connection timed out(110) while trying to reach CLDB: <cldb-host>:7223​
  The credentials are invalid.	Using maprlogin to authenticate with username/password..​. Authentication failed. Invalid username/password.​
  The mount prefix does not exist.	Mount prefix /ezua does not exist, attempting to create it..​. mkdir: User ezua(user id 7000) has been denied access to create ezua​ Mount prefix /ezua does not exist and could not create it​

To access the log files, run the following command:

kubectl -n mgmt-cluster-ops-system logs op-installua-ezaf100

If you cannot resolve the issue using the information provided in the log files, contact HPE Support.