Hello, - call home functionality - https://github.com/ceph/ceph/pull/51538 closed as a false start - something like this is definitely useful but there is a concern with making it a manager module since that could fracture cluster data collection interfaces; also push vs pull/scrape model - suggestion to take what is there in the exporter and run a translation layer in a sidecar which could be managed by cephadm as a third-party container to provide tighter integration; that way any new metric is readily available to all users (David) - not everything can be/is suitable to be a metric though, there is a variety of data sources (inventory data, one-off reports, etc) - there is a desire to generate and push "SOS report" blobs when issues are encountered - technical discussion to continue in other forums - reef 18.1.0 RC - https://github.com/ceph/ceph/pull/51900 is in - gibba cluster was upgraded (twice?) last week, no blocker issues - expecting to publish in the next day or two - better tracking for kernel client backports - currently manual, on a best-effort basis, by posting/editing comments; there is no support structure in the ticket - even this is often missed; it's not clear to users which release the original patch landed in, let alone a backport - proposal to add version drop-downs to "Linux kernel client" sub-project and have it generate backport tickets the same way it's done for most other sub-projects under "Ceph" - only upstream LTS kernels are in scope for now - resolving backport tickets can be automated with a new action to watch for respective commits showing up in linux-stable.git repo Thanks, Ilya _______________________________________________ Dev mailing list -- dev@xxxxxxx To unsubscribe send an email to dev-leave@xxxxxxx
