[Pgpool-general] Cannot accept() new connection

Clodoaldo clodoaldo.pinto.neto at gmail.com
Fri Mar 21 11:27:11 UTC 2008


I'm just starting with pgpool. I was using permanent connections with
both php and python and was having lost connection problems. Then I
decided to try pgpool. It is much simpler to setup than permanent
connections in python and so i like it very much. The only feature I'm
interested in pgpool is the connection pooling.

Pgpool II 2.0.1, Fedora 6, postgresql 8.2.6.

pgpool is installed in a web server and serving connections to the
port 5234 which is the local port for a reverse ssh tunnel to the
remote postgresql server.

The problems i had with my own permanent connections appear to also
happen with pgpool so I'm not here to blame pgpool, but to try to
figure it out. I don't know but i wonder if the ssh tunnel has any
relation with the problems. The ssh tunnel should be totally
transparent, shouldn't it?

The tcp/ip connection sometimes goes down. If it does not stay down
for very long the ssh tunnel persists and when the connection is back
the tunnel is still there and still working. But the pgpool
connections are locked. This has happened twice in the last 5 days.

I suppose pgpool would be capable of handling such situations so i
must be doing something wrong or missing something and i would like to
fix it.

Regards, Clodoaldo Pinto Neto

This is the log for the two incidents:

2008-03-20 19:45:43 LOG:   pid 2550: starting health checking
2008-03-20 19:48:44 LOG:   pid 2550: starting health checking
2008-03-20 19:51:44 LOG:   pid 2550: starting health checking
2008-03-20 19:52:18 ERROR: pid 23203: pool_read2: EOF encountered with backend
2008-03-20 19:52:18 LOG:   pid 23203: notice_backend_error: 0 fail
over request from pid 23203
2008-03-20 19:52:18 ERROR: pid 23225: pool_read2: EOF encountered with backend
2008-03-20 19:52:18 LOG:   pid 23225: notice_backend_error: 0 fail
over request from pid 23225
2008-03-20 19:52:18 ERROR: pid 23251: pool_read2: EOF encountered with backend
2008-03-20 19:52:18 LOG:   pid 23251: notice_backend_error: 0 fail
over request from pid 23251
2008-03-20 19:52:18 ERROR: pid 23246: pool_read2: EOF encountered with backend
2008-03-20 19:52:18 LOG:   pid 23246: notice_backend_error: 0 fail
over request from pid 23246
2008-03-20 19:52:18 ERROR: pid 2550: health check failed during write.
host localhost at port 5434 is down. reason: Broken pipe. Perhaps
wrong health check user?
2008-03-20 19:52:18 LOG:   pid 2550: set 0 th backend down status
2008-03-20 19:52:18 LOG:   pid 2550: starting degeneration. shutdown
host localhost(5434)
2008-03-20 19:52:18 ERROR: pid 2550: failover_handler: no valid DB node found
2008-03-20 19:52:18 LOG:   pid 2550: failover done. shutdown host
localhost(5434)
2008-03-20 19:52:18 LOG:   pid 2550: failover: no backends are degenerated
2008-03-20 19:52:28 LOG:   pid 23322: Cannot accept() new connection.
all backends are down
2008-03-20 19:52:31 LOG:   pid 23300: Cannot accept() new connection.
all backends are down
2008-03-20 19:52:32 LOG:   pid 23320: Cannot accept() new connection.
all backends are down
2008-03-20 19:52:35 LOG:   pid 23318: Cannot accept() new connection.
all backends are down
2008-03-20 19:53:02 LOG:   pid 23312: Cannot accept() new connection.
all backends are down
2008-03-20 19:53:02 LOG:   pid 23304: Cannot accept() new connection.
all backends are down
2008-03-20 19:53:27 LOG:   pid 23302: Cannot accept() new connection.
all backends are down
2008-03-20 19:53:54 LOG:   pid 23316: Cannot accept() new connection.
all backends are down
2008-03-20 19:53:54 LOG:   pid 23314: Cannot accept() new connection.
all backends are down
2008-03-20 19:53:54 LOG:   pid 23306: Cannot accept() new connection.
all backends are down
2008-03-20 19:53:58 LOG:   pid 23310: Cannot accept() new connection.
all backends are down
2008-03-20 19:54:31 LOG:   pid 23308: Cannot accept() new connection.
all backends are down
2008-03-20 19:55:06 LOG:   pid 23323: Cannot accept() new connection.
all backends are down
2008-03-20 19:55:18 LOG:   pid 2550: starting health checking
2008-03-20 19:55:19 LOG:   pid 23321: Cannot accept() new connection.
all backends are down
2008-03-20 19:55:36 LOG:   pid 23317: Cannot accept() new connection.
all backends are down
2008-03-20 19:55:36 LOG:   pid 23319: Cannot accept() new connection.
all backends are down
2008-03-20 19:55:38 LOG:   pid 23305: Cannot accept() new connection.
all backends are down
2008-03-20 19:55:43 LOG:   pid 23311: Cannot accept() new connection.
all backends are down
2008-03-20 19:55:43 LOG:   pid 23315: Cannot accept() new connection.
all backends are down
2008-03-20 19:56:02 LOG:   pid 23307: Cannot accept() new connection.
all backends are down
2008-03-20 19:56:42 LOG:   pid 23313: Cannot accept() new connection.
all backends are down
2008-03-20 19:57:05 LOG:   pid 23309: Cannot accept() new connection.
all backends are down
2008-03-20 19:57:05 LOG:   pid 23301: Cannot accept() new connection.
all backends are down
2008-03-20 19:57:11 LOG:   pid 23299: Cannot accept() new connection.
all backends are down
2008-03-20 19:57:12 LOG:   pid 23297: Cannot accept() new connection.
all backends are down
2008-03-20 19:57:45 LOG:   pid 23296: Cannot accept() new connection.
all backends are down
2008-03-20 19:57:49 LOG:   pid 23295: Cannot accept() new connection.
all backends are down
2008-03-20 19:57:49 LOG:   pid 23294: Cannot accept() new connection.
all backends are down
2008-03-20 19:57:49 LOG:   pid 23293: Cannot accept() new connection.
all backends are down
2008-03-20 19:57:49 LOG:   pid 23292: Cannot accept() new connection.
all backends are down
2008-03-20 19:58:18 LOG:   pid 2550: starting health checking
2008-03-20 20:01:18 LOG:   pid 2550: starting health checking
2008-03-20 20:04:18 LOG:   pid 2550: starting health checking
2008-03-20 20:07:18 LOG:   pid 2550: starting health checking
2008-03-20 20:10:18 LOG:   pid 2550: starting health checking
2008-03-20 21:54:44 LOG:   pid 2550: received smart shutdown request
2008-03-20 21:55:18 LOG:   pid 25682: pgpool successfully started
2008-03-20 21:55:24 LOG:   pid 25682: starting health checking


2008-03-21 07:45:48 LOG:   pid 27832: starting health checking
2008-03-21 07:50:48 LOG:   pid 27832: starting health checking
2008-03-21 07:54:55 ERROR: pid 2142: pool_read2: EOF encountered with backend
2008-03-21 07:54:55 LOG:   pid 2142: notice_backend_error: 0 fail over
request from pid 2142
2008-03-21 07:54:55 ERROR: pid 2239: pool_read2: EOF encountered with backend
2008-03-21 07:54:55 LOG:   pid 2239: notice_backend_error: 0 fail over
request from pid 2239
2008-03-21 07:54:55 LOG:   pid 27832: starting degeneration. shutdown
host localhost(5434)
2008-03-21 07:54:55 ERROR: pid 27832: failover_handler: no valid DB node found
2008-03-21 07:54:55 LOG:   pid 27832: failover done. shutdown host
localhost(5434)
2008-03-21 07:55:50 LOG:   pid 27832: starting health checking
2008-03-21 08:00:50 LOG:   pid 27832: starting health checking
2008-03-21 08:05:50 LOG:   pid 27832: starting health checking
2008-03-21 08:10:50 LOG:   pid 27832: starting health checking
2008-03-21 08:15:50 LOG:   pid 27832: starting health checking
2008-03-21 08:19:43 LOG:   pid 2257: Cannot accept() new connection.
all backends are down
2008-03-21 08:20:09 LOG:   pid 2266: Cannot accept() new connection.
all backends are down
2008-03-21 08:20:12 LOG:   pid 2323: Cannot accept() new connection.
all backends are down
2008-03-21 08:20:18 LOG:   pid 2268: Cannot accept() new connection.
all backends are down
2008-03-21 08:20:19 LOG:   pid 2325: Cannot accept() new connection.
all backends are down
2008-03-21 08:20:47 LOG:   pid 2326: Cannot accept() new connection.
all backends are down
2008-03-21 08:20:50 LOG:   pid 27832: starting health checking
2008-03-21 08:21:04 LOG:   pid 2330: Cannot accept() new connection.
all backends are down
2008-03-21 08:21:23 LOG:   pid 2349: Cannot accept() new connection.
all backends are down
2008-03-21 08:21:23 LOG:   pid 2264: Cannot accept() new connection.
all backends are down
2008-03-21 08:21:24 LOG:   pid 2259: Cannot accept() new connection.
all backends are down
2008-03-21 08:21:30 LOG:   pid 2258: Cannot accept() new connection.
all backends are down
2008-03-21 08:21:35 LOG:   pid 2267: Cannot accept() new connection.
all backends are down
2008-03-21 08:21:37 LOG:   pid 2265: Cannot accept() new connection.
all backends are down
2008-03-21 08:21:38 LOG:   pid 2262: Cannot accept() new connection.
all backends are down
2008-03-21 08:21:40 LOG:   pid 2260: Cannot accept() new connection.
all backends are down
2008-03-21 08:21:40 LOG:   pid 2261: Cannot accept() new connection.
all backends are down
2008-03-21 08:21:42 LOG:   pid 2263: Cannot accept() new connection.
all backends are down
2008-03-21 09:35:28 LOG:   pid 27832: received smart shutdown request
2008-03-21 09:35:29 LOG:   pid 4316: pgpool successfully started
2008-03-21 09:35:29 LOG:   pid 4316: starting health checking


More information about the Pgpool-general mailing list