[LU-14277] any create blocked due any OST fail

• Type: Bug
• Resolution: Fixed
• Priority: Blocker
• Fix Version/s: Lustre 2.15.0
• Severity: 3
• Assignee: Alexey Lyashkov
• Reporter: Alexey Lyashkov

Description

multiple MDT threads blocked by one declare_create call due OST fail.
rwsem_down_write_failed
call_rwsem_down_write_failed
lod_qos_prep_create
lod_declare_striped_object
lod_declare_object_create
mdd_declare_object_create_internal
mdd_declare_create
mdd_create
mdo_create
mdt_reint_open
mdt_reint_rec
mdt_reint_internal
mdt_intent_reint
mdt_intent_policy
ldlm_lock_enqueue
ldlm_handle_enqueue0
tgt_enqueue
tgt_request_handle
ptlrpc_main
kthread
vs
osp_precreate_reserve
osp_declare_object_create
lod_qos_declare_object_on
lod_alloc_rr.clone.2
lod_qos_prep_create
lod_declare_striped_object
lod_declare_object_create
mdd_declare_object_create_internal
mdd_declare_create
mdd_create
mdo_create
mdt_reint_open
mdt_reint_rec
mdt_reint_internal
mdt_intent_reint
mdt_intent_policy
ldlm_lock_enqueue
ldlm_handle_enqueue0
tgt_enqueue
tgt_request_handle
ptlrpc_main
kthread
kernel_thread

Issue Links

• duplicates: LU-13073 Multiple MDS deadlocks (in lod_qos_prep_create) after OSS crash
• is related to: LU-7853 Fixes bitfield in lod qos code
• is related to: LU-15393 object allocation when OST is lost