View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000002 | Pgpool-II | Bug | public | 2012-02-06 22:39 | 2012-02-14 16:29 |
| Reporter | tuomas | Assigned To | t-ishii | ||
| Priority | normal | Severity | major | Reproducibility | always |
| Status | resolved | Resolution | open | ||
| OS | Ubuntu | OS Version | 11.04 | ||
| Summary | 0000002: Client connection hangs in daemon mode | ||||
| Description | The issues happens at least with md5 pool authentication enabled pgpool version 3.1.2 postgresql 9.1 pg_hba.conf: host all all 0.0.0.0/0 md5 pool_hba.conf: host all all 0.0.0.0/0 md5 pgpool.conf: streaming master-slave configuration is used. ssl = off in pgpool and server. enable_pool_hba = on authentication_timeout = 5 user/password pair generated by pg_md5 is in pool_passwd. when trying to connect to pgpool, by i.e. $ psql --host=127.0.0.1 --user=username -d template1 The "pgpool: accept connection" and the psql process just hang up forever. Nothing happens, no errors in pgpool or stdout/err. In postgresql server log is an error: 2012-02-06 13:15:38.636 UTC LOG: connection received: host=10.x.x.x port=45207 2012-02-06 13:15:38.639 UTC LOG: could not receive data from client: Connection reset by peer 2012-02-06 13:15:38.639 UTC LOG: unexpected EOF within message length word In debug mode I get the following to syslog: Feb 6 13:33:57 qa01 pgpool[18845]: I am 18845 accept fd 6 Feb 6 13:33:57 qa01 pgpool[18845]: connection received: host=localhost port=46606 Feb 6 13:33:57 qa01 pgpool[18845]: Protocol Major: 1234 Minor: 5679 database: user: Feb 6 13:33:57 qa01 pgpool[18845]: SSLRequest from client Feb 6 13:33:57 qa01 pgpool[18845]: pool_ssl: SSL requested but SSL support is not available Feb 6 13:33:57 qa01 pgpool[18845]: read_startup_packet: application_name: psql Feb 6 13:33:57 qa01 pgpool[18845]: Protocol Major: 3 Minor: 0 database: yyy user: xxx Nothing after that. Strangely enough, everything seem to work fine if i run pgpool in foreground mode with -n. | ||||
| Tags | No tags attached. | ||||
|
|
This seems to happen also in foreground mode when for example, there are no backends available or when pgpool has received smart shutdown request but doesn't shutdown because of the other bug i reported. Clients just hang there forever. There are possibly some other cases where this might happen, too. |
|
|
Fix committed to master and 3.1 STABLE. So fix will appear in pgpool-II 3.2 and 3.1.3. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2012-02-06 22:39 | tuomas | New Issue | |
| 2012-02-11 19:23 | tuomas | Note Added: 0000004 | |
| 2012-02-14 16:20 | t-ishii | Assigned To | => t-ishii |
| 2012-02-14 16:20 | t-ishii | Status | new => assigned |
| 2012-02-14 16:29 | t-ishii | Note Added: 0000006 | |
| 2012-02-14 16:29 | t-ishii | Status | assigned => resolved |
