I have a 9.5alpha1 cluster which is locked up. All the user back ends seem to be waiting on semop, eventually on WALInsertLockAcquire.
Is there a way to use gdb to figure out who holds the lock they are waiting for? It is compiled with both debug and cassert. I am hoping someone can give me recipe similar to the one for Examining backend memory use in https://wiki.postgresql.org/wiki/Developer_FAQ example backtrace: #0 0x0000003dcb6eaf27 in semop () from /lib64/libc.so.6 #1 0x000000000067190f in PGSemaphoreLock (sema=0x7f28a98b9468) at pg_sema.c:387 #2 0x00000000006d4b0c in LWLockAcquireCommon (l=0x7f28a0e6d600, valptr=0x7f28a0e6d618, val=0) at lwlock.c:1042 #3 LWLockAcquireWithVar (l=0x7f28a0e6d600, valptr=0x7f28a0e6d618, val=0) at lwlock.c:916 #4 0x00000000004f3c4f in WALInsertLockAcquire (rdata=0xc5c130, fpw_lsn=0) at xlog.c:1411 #5 XLogInsertRecord (rdata=0xc5c130, fpw_lsn=0) at xlog.c:948 #6 0x00000000004f7aac in XLogInsert (rmid=13 '\r', info=32 ' ') at xloginsert.c:453 #7 0x000000000047e0b0 in ginPlaceToPage (btree=0x7fffca9263e0, stack=0x2c94ff8, insertdata=<value optimized out>, updateblkno=<value optimized out>, childbuf=0, buildStats=0x0) at ginbtree.c:418 #8 0x000000000047f3ad in ginInsertValue (btree=0x7fffca9263e0, stack=0x2c94ff8, insertdata=0x7fffca926460, buildStats=0x0) at ginbtree.c:748 #9 0x0000000000475c8b in ginEntryInsert (ginstate=0x7fffca9267e0, attnum=29784, key=1, category=<value optimized out>, items=0x7f28a0c7b458, nitem=47, buildStats=0x0) at gininsert.c:234 #10 0x0000000000485ecc in ginInsertCleanup (ginstate=0x7fffca9267e0, vac_delay=<value optimized out>, stats=0x0) at ginfast.c:843 #11 0x0000000000487059 in ginHeapTupleFastInsert (ginstate=0x7fffca9267e0, collector=<value optimized out>) at ginfast.c:436 #12 0x00000000004760fa in gininsert (fcinfo=<value optimized out>) at gininsert.c:531 Cheers, Jeff
