Uploaded image for project: 'WiredTiger'
  1. WiredTiger
  2. WT-1451

test/fops eviction hang

    XMLWordPrintable

    Details

    • Type: Task
    • Status: Closed
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: None
    • Labels:
      None

      Description

      test/fops has been failing in Jenkins because it is hanging. I can only reproduce this on tinderbox, not any other machine I've tried. This is not always reproducible. Often running again Jenkins passes.

      Eviction is hung, and it is holding the dhandle lock so everything else is blocked.

            3 __lll_lock_wait,_L_lock_812,__GI___pthread_mutex_lock,__wt_spin_lock,__session_checkpoint,obj_checkpoint,fop,start_thread,clone
            2 __lll_lock_wait,_L_lock_812,__GI___pthread_mutex_lock,__wt_spin_lock,__wt_curfile_open,__wt_open_cursor,__wt_curtable_open,__wt_open_cursor,__session_open_cursor,obj_bulk_unique,fop,start_thread,clone
            2 
            1 pthread_join,fop_start,main
            1 pthread_cond_timedwait@@GLIBC_2.3.2,__wt_cond_wait,__sweep_server,start_thread,clone
            1 __lll_lock_wait,_L_lock_812,__GI___pthread_mutex_lock,__wt_spin_lock,__wt_session_get_btree,__wt_session_get_btree_ckpt,__wt_curfile_open,__wt_open_cursor,__wt_curtable_open,__wt_open_cursor,__session_open_cursor,obj_cursor,fop,start_thread,clone
            1 __lll_lock_wait,_L_lock_812,__GI___pthread_mutex_lock,__wt_spin_lock,__wt_session_get_btree,__wt_session_get_btree_ckpt,__wt_curfile_open,__wt_open_cursor,__wt_curtable_open,__wt_open_cursor,__session_open_cursor,obj_bulk_unique,fop,start_thread,clone
            1 __lll_lock_wait,_L_lock_812,__GI___pthread_mutex_lock,__wt_spin_lock,__wt_session_get_btree,__wt_metadata_open,__schema_add_table,__wt_schema_get_table,__create_table,__wt_schema_create,__session_create,obj_bulk,fop,start_thread,clone
            1 __lll_lock_wait,_L_lock_812,__GI___pthread_mutex_lock,__wt_spin_lock,__wt_curfile_open,__wt_open_cursor,__wt_curtable_open,__wt_open_cursor,__session_open_cursor,obj_bulk,fop,start_thread,clone
            1 __lll_lock_wait,_L_lock_812,__GI___pthread_mutex_lock,__wt_spin_lock,__session_create,obj_bulk,fop,start_thread,clone
            1 __evict_walk,__evict_lru_walk,__evict_pass,__evict_server,start_thread,clone
      

      Here's the stack

      Thread 13 (Thread 0x7f7e24a3a700 (LWP 26032)):
      #0  0x00007f7e24ac9a0d in __evict_walk (flags=<optimized out>, session=0x151d7f0) at ../src/evict/evict_lru.c:901
      WT-1  __evict_lru_walk (flags=<optimized out>, session=0x151d7f0) at ../src/evict/evict_lru.c:725
      WT-2  __evict_pass (session=0x151d7f0) at ../src/evict/evict_lru.c:484
      WT-3  __evict_server (arg=0x151d7f0) at ../src/evict/evict_lru.c:163
      

      Putting breakpoints at key lines in *evict_walk shows that there are four open files and every single one of them has WT_BTREE_NO_EVICTION set. Therefore *evict_walk never returns anything to evict.

        Attachments

          Issue Links

            Activity

              People

              • Assignee:
                Unassigned
                Reporter:
                sue.loverso Sue LoVerso
              • Votes:
                0 Vote for this issue
                Watchers:
                1 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved: