Pivotal Knowledge Base

Follow

gpcmdr --setup fails with "Failure in talking to host smdw" error message

Environment 

Product Version
Pivotal Greenplum (GPDB) 4.3.x, 4.2.x
Greenplum command center 1.x

Symptom

The following error message is seen:

Creating instance schema in GPDB. Please wait ...
ERROR: extra data after last expected column (entry db basddwm1:5432 pid=30874)
DETAIL: External table fsmap, line 1 of execute:$GPPERFMONHOME/bin/gpgenfsmap.py 2> /dev/null || true: "20140829:15:16:26:030879 gpgenfsmap.py:default-[ERROR]:-Failure in talking to host smdw"

Cause

The issue is caused by the master having multiple file-spaces. Some file-spaces may be missing on the standby master and this can cause the "gpcmdr --setup" to fail.

Resolution

Manually create all the missing file-space directories on smdw, then run the gpcmdr --setup again.

Check the file-spaces used by the master and standby master: 

gpadmin=# select * from gp_segment_configuration where content=-1;
dbid | content | role | preferred_role | mode | status | port | hostname | address | replication_port | san_mounts
------+---------+------+----------------+------+--------+------+----------+---------+------------------+------------
1 |-1 | p| p| s| u| 5156 | mdw|  mdw
4 |-1 | m| m| s| u| 5156 | smdw smdw
(2 rows) gpadmin=# select * from pg_filespace_entry where fsedbid in (1,4); fsefsoid | fsedbid |fselocation ----------+---------+------------------------------- 3052 | 1 | /data/sigma_prod/sigma_prod-1 23584 | 1 | /data/test/sigma_prod-1 3052 | 4 | /data/sigma_prod/sigma_prod-1 23584 | 4 | /data/test/sigma_prod-1 (4 rows)  

For the standby master, (in this case, dbid=4) check whether the file spaces exist on server "smdw". If not, you have to create them manually and then test gpcmdr again: 

mkdir /data/test
mkdir /data/test/sigma_prod-1

Note: The standby master may not be in sync with the master. Please use the "gpinitstandby" command to reinitialize the standby master if you see the above behavior.

Comments

Powered by Zendesk