Details
Description
With the latest Forestdb commit, compaction does not complete after running for 12 hours. All other threads in benchmark test have completed except the thread handling compaction.
(gdb) t a a bt
Thread 4 (Thread 0x7f0c45a8b700 (LWP 30132)):
#0 0x0000003b3ba0b98e in pthread_cond_timedwait@@GLIBC_2.3.2 () from /lib64/libpthread.so.0
#1 0x00007f0c45cb9e47 in compactor_thread (voidargs=<value optimized out>) at /root/forestdb/src/compactor.cc:416
#2 0x0000003b3ba079d1 in start_thread () from /lib64/libpthread.so.0
#3 0x0000003b3b6e89dd in clone () from /lib64/libc.so.6
Thread 3 (Thread 0x7f0c4508a700 (LWP 30338)):
#0 0x0000003b3ba0822d in pthread_join () from /lib64/libpthread.so.0
#1 0x0000000000402c8f in do_incremental_writes (arg=0x192a730) at forestdb_workload.cc:375
#2 0x0000003b3ba079d1 in start_thread () from /lib64/libpthread.so.0
#3 0x0000003b3b6e89dd in clone () from /lib64/libc.so.6
Thread 2 (Thread 0x7f0c3d7fb700 (LWP 30342)):
#0 avl_search (tree=0x7f0c2800c270, node=0x7f0c3d7fa4e0, func=0x7f0c45cd3f20 <_hash_cmp_wrap(avl_node*, avl_node*, void*)>)
at /root/forestdb/src/avltree.cc:323
#1 0x00007f0c45cd4115 in _hash_find (ht=<value optimized out>, e=<value optimized out>, hash_val=<value optimized out>)
at /root/forestdb/src/hash.cc:88
#2 hash_find_by_hash_val (ht=<value optimized out>, e=<value optimized out>, hash_val=<value optimized out>)
at /root/forestdb/src/hash.cc:119
#3 0x00007f0c45cd5927 in wal_insert (txn=0x7f0c280020d8, file=0x7f0c28002000, doc=0x7f0c3d7fa740, offset=121888742168,
is_compactor=1, mmap_alloc=0) at /root/forestdb/src/wal.cc:232
#4 0x00007f0c45cc87da in __fdb_flush_doc_batch (handle=0x7f0c280008f0, new_file=0x7f0c28002000,
new_trie=<value optimized out>, new_seqtree=<value optimized out>, new_dhandle=0x7f0c28022a00,
new_bhandle=<value optimized out>, got_lock=false, new_idtree=0x0) at /root/forestdb/src/forestdb.cc:3710
#5 _fdb_compact_move_docs_seq (handle=0x7f0c280008f0, new_file=0x7f0c28002000, new_trie=<value optimized out>,
new_seqtree=<value optimized out>, new_dhandle=0x7f0c28022a00, new_bhandle=<value optimized out>, got_lock=false,
new_idtree=0x0) at /root/forestdb/src/forestdb.cc:3871
#6 0x00007f0c45cc8e8f in _fdb_compact_file (handle=0x7f0c280008f0, new_file=0x7f0c28002000, new_bhandle=0x7f0c28022990,
new_dhandle=0x7f0c28022a00, new_trie=0x7f0c28023c10, new_seqtrie=0x7f0c28023d20, new_seqtree=0x0)
at /root/forestdb/src/forestdb.cc:4392
#7 0x00007f0c45cca649 in fdb_compact_file (fhandle=<value optimized out>, new_filename=<value optimized out>,
in_place_compaction=<value optimized out>) at /root/forestdb/src/forestdb.cc:4303
#8 0x00007f0c45cca7c4 in fdb_compact (fhandle=0x7f0c280008c0, new_filename=<value optimized out>)
at /root/forestdb/src/forestdb.cc:4463
--Type <return> to continue, or q <return> to quit--
#9 0x00000000004034ab in compactor (arg=0x7f0c38008e50) at forestdb_workload.cc:525
#10 0x0000003b3ba079d1 in start_thread () from /lib64/libpthread.so.0
#11 0x0000003b3b6e89dd in clone () from /lib64/libc.so.6
Thread 1 (Thread 0x7f0c45a8d720 (LWP 30131)):
#0 0x0000003b3ba0822d in pthread_join () from /lib64/libpthread.so.0
#1 0x0000000000403a8c in do_incremental_load (num_wthreads=1, num_rthreads=2) at forestdb_workload.cc:594
#2 0x0000000000403eba in main (argc=1, args=0x7fffaeb766f8) at forestdb_workload.cc:647
Attachments
Issue Links
- relates to
-
MB-14060 Investigating compaction's impact on incremental workload throughput.
- Closed