[check_postgres] interaction between sudo, pg_controldata, and check_postgres (action=checkpoint)

[James Skaggs]

Greg -

I appreciate check_postgres and use it a lot.

I think I found an interaction between sudo, pg_controldata, and
check_postgres.  There is a proposed fix at the bottom.

check_postgres: Version 2.19.0
postgresql:  8.4.15

[nagios at cyclonus ~]$ PGBINDIR=/usr/pgsql-8.4/bin/
[nagios at cyclonus ~]$ export PGBINDIR
[nagios at cyclonus ~]$ sudo -u postgres
/usr/bin/check_postgres.pl--assume-standby-mode --datadir
/pgdata/coverity/data/ --action=checkpoint
*ERROR: Failed to run pg_controldata - is it the correct version?*

In the perl debugger, we see extraneous output from pg_controldata....BTW,
his does not happen when run directly from the "postgres" user.

  DB<3> p $COM
/usr/pgsql-8.4/bin//pg_controldata "/pgdata/coverity/data/"
3181:       if ($res !~ /^pg_control.+\d+/) {
  DB<7> p $res
*could not change directory to "/var/log/nagios"*
pg_control version number:            843
Catalog version number:               200904091
Database system identifier:           5805897768012068846
Database cluster state:               in archive recovery
pg_control last modified:             Mon 14 Jan 2013 03:54:30 PM MST
Latest checkpoint location:           F0/109D8900
Prior checkpoint location:            F0/74AE670
Latest checkpoint's REDO location:    F0/ABBED08
Latest checkpoint's TimeLineID:       1
Latest checkpoint's NextXID:          0/121770964
Latest checkpoint's NextOID:          426392
Latest checkpoint's NextMultiXactId:  4420942
Latest checkpoint's NextMultiOffset:  11062935
Time of latest checkpoint:            Mon 14 Jan 2013 03:46:22 PM MST
Minimum recovery ending location:     F0/2BFFFFD0
Maximum data alignment:               8
Database block size:                  8192
Blocks per segment of large relation: 131072
WAL block size:                       8192
Bytes per WAL segment:                16777216
Maximum length of identifiers:        64
Maximum columns in an index:          32
Maximum size of a TOAST chunk:        1996
Date/time type storage:               64-bit integers
Float4 argument passing:              by value
Float8 argument passing:              by value

  DB<8> v
3178:       if ($res =~ /WARNING: Calculated CRC checksum/) {
3179:           ndie msg('checkpoint-badver', $dir);
3180        }
3181==>     if ($res !~ /^pg_control.+\d+/) {
3182:           ndie msg('checkpoint-badver2');
3183        }
3184

The extra output breaks the regex at line 3181.  Adding a "m" option lets
it match at the beginning of any line, not just the first.

3181c3181
<     if ($res !~ /^pg_control.+\d+/) {
---
>     if ($res !~ /^pg_control.+\d+/*m*) {

BTW, using sudo had it's own difficulties....mostly about sudo requiring a
tty by default, something that that nrpe does not provide.

diff ... /etc/sudoers
24a25
> Cmnd_Alias    NAGIOS_CMDS = /usr/bin/check_postgres.pl,
/var/lib/pgsql/check_posgres_checkpoint, /usr/pgsql-8.4/bin/pg_controldata
59a61,62
> # This is so nagios can use sudo to posgress to check the database
> *Defaults:nagios !requiretty*
114a118
> nagios  ALL=(postgres)  NOPASSWD: NAGIOS_CMDS

Regards,

James
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.endcrypt.com/pipermail/check_postgres/attachments/20130205/70ee04df/attachment.html>