[LU-6361] LFSCK 4: improve LFSCK performance Created: 13/Mar/15  Updated: 05/Feb/19  Resolved: 30/May/15

Status: Resolved
Project: Lustre
Component/s: None
Affects Version/s: None
Fix Version/s: Lustre 2.8.0

Type: Improvement Priority: Major
Reporter: nasf (Inactive) Assignee: nasf (Inactive)
Resolution: Fixed Votes: 0
Labels: None

Attachments: Microsoft Word LFSCK_4_test_results_201505.xlsx    
Sub-Tasks:
Key
Summary
Type
Status
Assignee
LU-1452 LFSCK 4: Improve inode iteration for ... Technical task Resolved nasf  
LU-1453 LFSCK 4: Improve OI scrub trigger str... Technical task Resolved nasf  
LU-5820 LFSCK 4: Record linkEA verification h... Technical task Resolved nasf  
LU-5682 LFSCK 4: optimize ldlm lock used by L... Technical task Resolved nasf  
LU-6316 LFSCK 4: namespace LFSCK can skip dot... Technical task Resolved nasf  
LU-6317 LFSCK 4: namespace LFSCK counts the o... Technical task Resolved nasf  
LU-6343 LFSCK 4: locate object only when nece... Technical task Resolved nasf  
LU-6350 LFSCK 4: lock the MDT-object based on... Technical task Resolved nasf  
LU-6322 LFSCK 4: show LFSCK started/completed... Technical task Resolved nasf  
LU-6177 LFSCK 4: namespace LFSCK scalability Technical task Resolved nasf  
LU-6351 LFSCK MDS crash: unable to handle ker... Technical task Resolved nasf  
Severity: 3
Rank (Obsolete): 9223372036854775807

 Description   

It is used for tracing all LFSCK 4 tasks.



 Comments   
Comment by nasf (Inactive) [ 01/May/15 ]

The patch for measure LFSCK 4 performance:
http://review.whamcloud.com/#/c/12109/

Comment by Gerrit Updater [ 30/May/15 ]

Oleg Drokin (oleg.drokin@intel.com) merged in patch http://review.whamcloud.com/12109/
Subject: LU-6361 lfsck: measure LFSCK 4 performance
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: db2110a7af734f42a3a0b425feade8498c309895

Comment by Peter Jones [ 30/May/15 ]

Landed for 2.8

Comment by John Bent (Inactive) [ 04/Feb/19 ]

Hello, has anyone actually used this awesome sounding patch to measure performance?  We have an RFP that wants 1B namespace entries to be fsck'd in 12 hours.  Anyone have any ideas what a reasonable lfsck rate is?  Thanks!

 

Comment by Andreas Dilger [ 05/Feb/19 ]

Attached are the test results comparing LFSCK phase 3 (DNE support) with LFSCK phase 4 (performance tuning), measured when this work was competed in 2015.

Generated at Sat Feb 10 01:59:32 UTC 2024 using Jira 9.4.14#940014-sha1:734e6822bbf0d45eff9af51f82432957f73aa32c.