mirror of
https://github.com/gravitl/netmaker.git
synced 2026-04-22 16:07:11 +08:00
Abhishek Kondur
b97fbc69db
* fix(go): set persistent keep alive when registering host using sso; * fix(go): run posture check violations on delete; * fix(go): upsert node on approving pending host; * fix(go): resolve concurrency issues during group delete cleanup; * fix(go): update doc links; * fix(go): add created and updated fields to host; * fix(go): skip delete and update superadmin on sync users; * fix(go): use conn directly for now; * fix(go): remove acl for idp groups; * fix(go): quote fields; * fix(go): use filters with count; * feat(go): add a search query; * fix(go): cleanup acls; * fix(go): review fixes; * fix(go): remove additional loop; * fix(go): fix * v1.5.1: separate out idp sync and reset signals for HA * v1.5.1: add grps with name for logging * v1.5.1: clear posture check violations when all checks are deleted * v1.5.1: set static when default host * v1.5.1: fix db status check * rm set max conns * v1.5.1: reset auto assigned gw when disconnected * fix(go): skip global network admin and user groups when splitting; * v1.5.1: fix update node call from client * fix(go): separate out migration from normal usage; * fix(go): skip default groups; * fix(go): create policies for existing groups on network create; * fix(go): skip fatal log on clickhouse conn; * fix(go): add posture check cleanup; * NM-288: populate relevant name for acl types for UI * NM-288: populate grp names for posture check apis * NM-228: add network grps api * NM-288: add network users api * now check each group's NetworkRoles for either the specific network ID or schema.AllNetworks (all_networks) * NM-288: check and unassign auto gw when node is disconnected from cli * NM-288: optimise network users api call * NM-288: block auto assign when set to use inet gw * NM-288: pass all network scoped get apis for platform users * NM-288: fix orphan nodes on bulk host deletion and refactor node cleanup - Extract cleanupNodeReferences() from DeleteNode to handle best-effort cleanup of relay, internet gw, failover, nameserver, ACL, and egress refs - DisassociateAllNodesFromHost now calls cleanupNodeReferences + DeleteNodeByID directly, bypassing redundant host-association updates that could fail and leave nodes orphaned - Track failed node deletions instead of unconditionally clearing host.Nodes; return error if any nodes couldn't be deleted to prevent host removal - Fix DeleteNode returning error even when node was successfully deleted in the host-not-found path * NM-288: cleanupNodeReferences(node) now only runs in the hard-delete path (when purge=true or alreadyDeleted=true), so it won't execute twice — once on soft-delete and again when the zombie manager hard-deletes. * NM-288: add orphan deletion log * DisassociateAllNodesFromHost now always returns nil, logging warnings for any partial failures (upsert errors or undeletable nodes). This allows RemoveHost to always proceed to h.Delete() --------- Co-authored-by: VishalDalwadi <dalwadivishal26@gmail.com> Co-authored-by: Vishal Dalwadi <51291657+VishalDalwadi@users.noreply.github.com>