Data Fabric CLDB unable to start

[SergioMarquez]

Hi, I've tried to connect with an "old" fabric cluster, apparently CLDB is down and tried to start it with the command "mapr-cldb start", but a apparently inner process "pullcentralconfig" fails and subsequently CLDB doesnt start, I've tried this because is not possible to generate a new ticket (for sure previous one has expired).

DF Version is 7.1 any idea?

Thank you!

[Dave Olker]

It's not recommended to start any EDF services directly via the initscripts.  The recommended approach is to use Warden to manage all the services.

I would try the following:

# systemctl stop mapr-warden
# systemctl start mapr-warden

Then check the /opt/mapr/pid directory to see if services are starting.  You can also monitor the Warden process with "systemctl status mapr-warden" to ensure it is active.  If Warden will not start, it may be due to Zookeeper not running or not having formed a quorum.  You can use similar systemctl stop/start/status commands to make sure mapr-zookeeper is also running correctly.

[SergioMarquez]

Thank you Dave, I've tried re-start of Warden, but apparently there's no similar process for zookeeper since I receive next reply:

"Unit mapr-zookeeper.service could not be found"

So, Im not sure is there another way to start it.

[Dave Olker · Accepted solution]

That tells me this was a multi-node cluster and zookeeper was installed on a different set of cluster node(s).  Was this cluster a multi-node cluster?  Do you have access to all the nodes in the original cluster or only this node?

[SergioMarquez]

You are right,  I've switched to the main node and tried Warden restart, apparently CLDB is up and running now (I've verified with JPS command)

Thanks a lot!

[support_s]

System recommended content:

1. HPE Ezmeral Data Fabric 7.1 Documentation | Known Issues (Release 7.1.0)

Please click on "Thumbs Up/Kudo" icon to give a "Kudo".

Thank you for being a HPE valuable community member.