SnapMirror Replication between the StorageGRID and ONTAP Fails with “IdleTransfer failed. (Failure to perform I/O with the capacity tier)” Error
Applies to
- StorageGRID
- ONTAP
- Veritas Backup
Issue
SnapMirror replication intermittently fails for multiple ONTAP volumes configured with FabricPool tiering to StorageGRID object store. The following error is observed:
IdleTransfer failed. (Failure to perform I/O with the capacity tier. Retry the operation)
Additional log output from StorageGRID gateway node:
2026/01/06 14:19:59 [warn] 1896053#1896053: *13880127723 [lua] sgbalancer.lua:647: on_balance_phase(): Connection to a615-1933259d9b9c failed with code 502 in worker 59 while connecting to upstream, client: <client IP>, server: _, request: "DELETE <Path> HTTP/1.1", upstream: "https://<storagenode ip>:18086/<storage node details>", host: "<Host details>:9021"
- SnapMirror relationships are marked as UNHEALTHY.
- Manual SnapMirror updates sometimes succeed, but automated replication fails.
- StorageGRID logs show multiple “connection refused” errors (
error code 111) and 502 gateway errors.
