Pivotal Greenplum Database segments on a Solaris machine fail to start and the segment logs report the following error message:
bash-3.00$ less gpdb-2014-01-06_194406.csv 2014-01-06 19:44:06.542036 IST,,,p27499,th1,,,,0,,,seg-1,,,,,"LOG","XX000","could not bind socket for statistics collector: Cannot assign requested address (pgstat.c:390)",,,,,,,0,,"pgstat.c",390, 2014-01-06 19:44:06.542233 IST,,,p27499,th1,,,,0,,,seg-1,,,,,"LOG","00000","trying another address for the statistics collector",,,,,,,0,,"pgstat.c",369, 2014-01-06 19:44:06.545910 IST,,,p27503,th1,,,,0,,,seg-1,,,,,"LOG","00000","database system was interrupted at 2014-01-06 19:42:30 IST",,,,,,,0,,"xlog.c",6338, 2014-01-06 19:44:06.546363 IST,,,p27503,th1,,,,0,,,seg-1,,,,,"LOG","00000","checkpoint record is at 0/8B9A60",,,,,,,0,,"xlog.c",6415, 2014-01-06 19:44:06.546627 IST,,,p27503,th1,,,,0,,,seg-1,,,,,"LOG","00000","redo record is at 0/8B9A60; undo record is at 0/0; shutdown TRUE",,,,,,,0,,"xlog.c",6524, 2014-01-06 19:44:06.546866 IST,,,p27503,th1,,,,0,,,seg-1,,,,,"LOG","00000","next transaction ID: 0/903; next OID: 10901",,,,,,,0,,"xlog.c",6528,
The exact reason is unknown but it might be related to the SolarisĀ configuration.
Try the workaround below to bring the database back up again:
::1 localhost
" from /etc/inet/ipnodes
on all segment hosts.