[LU-15712] lru_resize gets stuck with low slv value

• Type: Bug
• Resolution: Unresolved
• Priority: Major
• Affects Version/s: Lustre 2.12.8, Lustre 2.15.0
• Severity: 3
• Assignee: WC Triage
• Reporter: Oleg Drokin

Description

A client has observed a condition where SLV was stuck at 1 for prolonged period of time for no apparent reason. This lead to all clients not caching any locks for this service.

Since SLV is calculated/recalculated all the time some of that logic might be flawed I guess.

Captured values:

ldlm.namespaces.filter-xxx-OST0000_UUID.pool.cancel_rate=0
ldlm.namespaces.filter-xxx-OST0000_UUID.pool.client_lock_volume=0
ldlm.namespaces.filter-xxx-OST0000_UUID.pool.grant_plan=1981373
ldlm.namespaces.filter-xxx-OST0000_UUID.pool.grant_rate=49
ldlm.namespaces.filter-xxx-OST0000_UUID.pool.grant_speed=49
ldlm.namespaces.filter-xxx-OST0000_UUID.pool.granted=33966334
ldlm.namespaces.filter-xxx-OST0000_UUID.pool.limit=1585099
ldlm.namespaces.filter-xxx-OST0000_UUID.pool.lock_volume_factor=100
ldlm.namespaces.filter-xxx-OST0000_UUID.pool.recalc_period=1
ldlm.namespaces.filter-xxx-OST0000_UUID.pool.recalc_time=0
ldlm.namespaces.filter-xxx-OST0000_UUID.pool.server_lock_volume=1
  SLV: 1
  CLV: 0
  LVF: 100
  GSP: 1%
  GP:  1981373
  GR:  49
  CR:  0
  GS:  49
  G:   33966334
  L:   1585099

ldlm.namespaces.filter-xxx-OST0001_UUID.pool.cancel_rate=0
ldlm.namespaces.filter-xxx-OST0001_UUID.pool.client_lock_volume=0
ldlm.namespaces.filter-xxx-OST0001_UUID.pool.grant_plan=1981375
ldlm.namespaces.filter-xxx-OST0001_UUID.pool.grant_rate=14
ldlm.namespaces.filter-xxx-OST0001_UUID.pool.grant_speed=14
ldlm.namespaces.filter-xxx-OST0001_UUID.pool.granted=23738424
ldlm.namespaces.filter-xxx-OST0001_UUID.pool.limit=1585100
ldlm.namespaces.filter-xxx-OST0001_UUID.pool.lock_volume_factor=100
ldlm.namespaces.filter-xxx-OST0001_UUID.pool.recalc_period=1
ldlm.namespaces.filter-xxx-OST0001_UUID.pool.recalc_time=0
ldlm.namespaces.filter-xxx-OST0001_UUID.pool.server_lock_volume=1
  SLV: 1
  CLV: 0
  LVF: 100
  GSP: 1%
  GP:  1981375
  GR:  14
  CR:  0
  GS:  14
  G:   23738424
  L:   1585100

Issue Links

• is related to: LU-7266 Fix LDLM pool Server Lock Volume to make LRUR working properly