> On Aug 1, 2025, at 9:33 AM, gagan tiwari <gagan.tiwari@xxxxxxxxxxxxxxxxxx> wrote: > > Thanks Anthony, > That did the trick! All OSDs are up now. Groovy. > > Verifying port 0.0.0.0:9100 ... > Cannot bind to IP 0.0.0.0 port 9100: [Errno 98] Address already in use Yeah, that sure sounds like you already have node_exporter running. Prometheus exporters generally have a default port on which they listen, usually but not always in the 9xxx range. There's an informal registry of these somewhere to help avoid collisions, but it's not enforced. Most likely this is either detritus from a previous deployment attempt, or perhaps your organization has a fleetwide node_exporter deployment that is conflicting. I tend to place a fleet node_exporter deployment on 9101 instead of 9100 to avoid instances deployed by other tools. It doesn't hurt to have more than one running. 9100 is also historically the port used by HP's JetDirect network printer interface. > ERROR: TCP Port(s) '0.0.0.0:9100' required for node-exporter already in use > [WRN] CEPHADM_FAILED_DAEMON: 1 failed cephadm daemon(s) > daemon node-exporter.ceph-mon2 on ceph-mon2 is in error state > > I think most likely its due to the leftovers of the previous > installation. How would I go about removing this cleanly and more > importantly, in a way that Ceph is aware of the change, therefore clearing > the warning. systemctl -a ceph orch ls ceph orch ps Look for orphaned orchestrator instances and kill, then disable, kill, and remove with systemctl if the ceph orch commands don't do the whole job. node_exporter doesn't have state as such, so there's nothing to preserve. _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |