Bug 579832

Fix in lvm2-cluster-2.02.42-5.el4_8.1.

Looking at the backtrace it seems that the patch in bz#559999 didn't fix the bug (which is not *that* surprising to me as it was committed before anyone tested it).

The symptoms are identical. A client processing thread is blocked on the condition variable at clvmd.c:1454

1448                    pthread_mutex_lock(&client->bits.localsock.mutex);
1449                    DEBUGLOG("Waiting to do post command - state = %d\n",
1450                             client->bits.localsock.state);
1451    
1452                    if (client->bits.localsock.state != POST_COMMAND &&
1453                        !client->bits.localsock.finished) {
1454                            pthread_cond_wait(&client->bits.localsock.cond, 
1455                                              &client->bits.localsock.mutex);
1456                    }
1457                    pthread_mutex_unlock(&client->bits.localsock.mutex);
1458    
1459                    DEBUGLOG("Got post command condition...\n");
1460    
1461                    /* POST function must always run, even if the client aborts */
1462                    status = 0;
1463                    do_post_command(client);

(gdb) p client->bits.localsock.finished
$3 = 0
(gdb) p client->bits.localsock.state
$4 = PRE_COMMAND
(gdb) p status
$5 = 0

I'd also like to see some clvmd debug logs for this incident too.