2669056 - Corruption on LFB index after REORG REBUILD on index - SAP ASE 16.0

• Latch Free B-tree (LFB) index gets corrupt when running REORG REBUILD on that index: 

reorg rebuild table_name LFB_index_name

• Errors reported after this message in the ASE errorlog:
    
     Beginning REORG REBUILD of index 'LFB_index_name' on table 'table_name' in database 'my_db'.
   
  can include 12314, 820, 872 or 11068 (any of these errors):
   
• Error: 12314, Severity: 21, State: 2
  Bad child page pointer found on page 1234567 of index with id 4 of table 'table_name' in database 'my_db'. Expected child page pointer to be 0 but found 456789.
  Check if the index is consistent using DBCC.
   
• Msg 820, Level 21, State 2:
  Attempt to dirty non-log buffer BUF=0x0x0000000303626000, (.....) cid=0 ('default data cache'), Pinned xdes=0x(nil), spid=0 which is in I/O.
   
• Msg 872, Level 20, State 1:
  Process 4321 is trying to get a exclusive latch on buffer 0x0x303626000 (dbid: 4, pageno: 1882928) while already holding a exclusive latch on the buffer.
  Error while undoing log row in database 'my_db'.  Rid pageid = 10563434; row num = 1.
   
• Msg 11068, Level 20, State 1:
  Transaction was found in the incorrect state of 'Command-attached'. The expected state was 'Done command-attached'.
  ASE is terminating this process.
   
• Database may fail to recover after seeing above error(s), reporting Error 3478 :

Error: 3478, Severity: 21, State: 1
During undo the page timestamp value is less than new timestamp from log. Page #=1882928, object id = 407537725, page timestamp=0002 33da7127.
Log new timestamp=0002 33da7167. Log record marker = (10563434, 1).
Error while undoing log row in database 'my_db'.  Rid pageid = 10563434; row num = 1.

• ASE data cache configured with a large IO pool.

• SAP Adaptive Server Enterprise (ASE) 16.0 SP02 or higher
• Latch Free B-tree (LFB) index feature

CR815811, CR#815811, 815811, latch_free_index , KBA , BC-SYB-ASE , Sybase ASE Database Platform (non Business Suite) , Bug Filed