I'm getting a db crash with "SYSTEM ERROR: Releasing multiplexed latch" in the .lg file. i saw a k-base article about this when the BI file filled the disk. That's not happening here - my B1 is only 128k and there's plenty of room.
What else causes his error?
This is enterprise 10.2b on redhat linux 7.
Here's a snippet:
[2017/05/17@12:02:13.716-0400] P-5258 T--1860133968 I WDOG 17: (5029) S
YSTEM ERROR: Releasing multiplexed latch. latchId: 3096224744692744
[2017/05/17@12:02:13.716-0400] P-5258 T--1860133968 I WDOG 5: (2522) U
ser 17 died holding 1 shared memory locks.
[2017/05/17@12:02:13.716-0400] P-5258 T--1860133968 I WDOG 5: (-----) S
ending signal 12 to user 0
[2017/05/17@12:02:13.716-0400] P-5258 T--1860133968 I WDOG 5: (-----) S
ending signal 12 to user 6
[2017/05/17@12:02:13.716-0400] P-5258 T--1860133968 I WDOG 5: (-----) S
ending signal 12 to user 7
[2017/05/17@12:02:13.716-0400] P-5258 T--1860133968 I WDOG 5: (-----) S
ending signal 12 to user 8
[2017/05/17@12:02:13.716-0400] P-5255 T--520912768 I BROKER 0: (15192) Th
e database will complete shutdown within approximately 60 seconds.
Thanks in advance -- Jim
On UNIX: kill -9
Or, on Windows, terminating a process with task manager. Or logging off the console when some background process shares your userid.
> Releasing multiplexed latch. latchId: 3096224744692744
3096224744692744 = 0x000B0000 000D5C08
It's a BUF latch family. Process was killed while it read a block in buffer pool.
Find the PID of user 17 in the db log - "User 17 died holding 1 shared memory locks"
As Tom said, usually this is from a process getting killed and not being able to back out its transaction.
If this happens regularly, you might be able to find a pattern.