CONTAP-802858: EMS reports sis.chkpoint.restore.failed
Issue
- Running a fixed version for CONTAP-302610
- A "sis.chkpoint.restore.failed" EMS event is logged against a volume with volume efficiency enabled, and the efficiency operation restarts from the beginning instead of resuming:
[NODE-02: dense_worker: sis.chkpoint.restore.failed:error]: SIS operation on the volume '/vol/<volume>@vserver:<uuid>' failed to restore from a previous checkpoint. Starting SIS operation from the beginning. - The corresponding volume efficiency log (sis.log) shows a "checkpoint restore failed" error, immediately followed by a new operation "start" on the same volume from any trigger (like "vault" or "schedule"), and the new operation beginning with a "Verify" phase:
example on a DP volume when snapmirror update starts:
Mon Sep 7 19:04:52 BST 2026 [Vserver UUID: <uuid>] /vol/<volume> [sid: 1784778504] Warning (Unable to restore the checkpoint from stage: Checking_pass1 with err: Checkpoint restore failed)
Mon Sep 7 19:05:03 BST 2026 [Vserver UUID: <uuid>] /vol/<volume> [sid: 1788804292] Begin (sis start vault)
Mon Sep 7 19:05:03 BST 2026 [Vserver UUID: <uuid>] /vol/<volume> [sid: 1788804292] Verify
example on RW volume when ADS throttles:
Sun Sep 6 01:00:23 BST 2026 [Vserver UUID: <uuid>] /vol/<volume> [sid: 1788605875] Info (ADS: Op is evicted to schedule clog ops.)
Sun Sep 6 01:00:23 BST 2026 [Vserver UUID: <uuid>] /vol/<volume> [sid: 0] Info (ADS: Automatic Dedupe operation is pausing because of throttling or incoming user op)
Sun Sep 6 01:00:25 BST 2026 [Vserver UUID: <uuid>] /vol/<volume> [sid: 0] Error (Operation was stopped to make resources available for a user request)
Sun Sep 6 01:00:28 BST 2026 [Vserver UUID: <uuid>] /vol/<volume> [sid: 1788652828] Begin (sis auto start)
Sun Sep 6 01:00:29 BST 2026 [Vserver UUID: <uuid>] /vol/<volume> [sid: 1788605875] Warning (Unable to restore the checkpoint from stage: Checking_pass2 with err: Checkpoint restore failed) - from ONTAP 9.16.1, the sktrace log reports
Old input bin <x> not present, like:
DENSE_INFO: [checkpoint.c:1525][dense_check_merge_sort_checkpoint]: Old input bin 0 not present
DENSE_INFO: [checkpoint.c:1525][dense_check_merge_sort_checkpoint]: Old input bin 1 not present
...
DENSE_INFO: [checkpoint.c:1525][dense_check_merge_sort_checkpoint]: Old input bin 1279 not present
DENSE_ERR: [checkpoint.c:1531][dense_check_merge_sort_checkpoint]: Unable to open dense metafile for /sort_verify/bin1280 with error Metafile does not exist
DENSE_ERR: [checkpoint.c:5791][dense_restore_ckpt_checking_p1]: Failed to restore from verify P1 ckpt on vol 1042 with error Checkpoint restore failed - The situation may show once, or repeatedly, may solve on its own or require human intervention to allow a "verify" phase to complete
