VMware Tanzu Greenplum

 View Only

 auto_explain is crashing greenplum segment hosts

Dolphin Madan's profile image
Dolphin Madan posted Feb 26, 2020 07:38 AM

I'm running a Greenplum on a single machine having both master and segments. I've enabled auto_explain on both master and segment hosts. I'm seeing a greenplum crash whenever i run a query that involves redistribute motion plan. Below are the details:

 

GreenPlum State:

[gpadmin@sg-madstand-88 ~]$ gpstate

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-Starting gpstate with args:

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-local Greenplum Version: 'postgres (Greenplum Database) 6.0.1 build commit:053a66ae19cd7301ec8c8910ed85ec2c20ad60cc'

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-master Greenplum Version: 'PostgreSQL 9.4.24 (Greenplum Database 6.0.1 build commit:053a66ae19cd7301ec8c8910ed85ec2c20ad60cc) on x86_64-unknown-linux-gnu, compiled by gcc (GCC) 6.4.0, 64-bit compiled on Oct 11 2019 18:47:35'

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-Obtaining Segment details from master...

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-Gathering data from segments...

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-Greenplum instance status summary

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-----------------------------------------------------

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Master instance                = Active

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Master standby                 = No master standby configured

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Total segment instance count from metadata   = 2

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-----------------------------------------------------

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Primary Segment Status

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-----------------------------------------------------

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Total primary segments             = 2

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Total primary segment valid (at master)    = 2

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Total primary segment failures (at master)   = 0

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Total number of postmaster.pid files missing  = 0

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Total number of postmaster.pid files found   = 2

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Total number of postmaster.pid PIDs missing  = 0

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Total number of postmaster.pid PIDs found   = 2

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Total number of /tmp lock files missing    = 0

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Total number of /tmp lock files found     = 2

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Total number postmaster processes missing   = 0

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Total number postmaster processes found    = 2

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-----------------------------------------------------

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Mirror Segment Status

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-----------------------------------------------------

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-  Mirrors not configured on this array

20200226:03:44:59:016864 gpstate:sg-madstand-88:gpadmin-[INFO]:-----------------------------------------------------

 

Table definition:

madan=# \dS+ mad;

             Table "public.mad"

 Column | Type  | Modifiers | Storage | Stats target | Description

--------+---------+-----------+---------+--------------+-------------

 id   | integer |      | plain  |       |

Distributed by: (id)

 

Query that is crashing:

 

[gpadmin@sg-madstand-88 ~]$ psql -d madan

psql (9.4.24)

Type "help" for help.

 

madan=# select gp_segment_id, count(*) from mad group by 1;

WARNING: bogus varno: 65001 (seg0 slice2 172.31.43.226:7000 pid=14069)

WARNING: bogus varno: 65001 (seg0 slice2 172.31.43.226:7000 pid=14069)

WARNING: bogus varno: 65001 (seg0 slice2 172.31.43.226:7000 pid=14069)

WARNING: bogus varno: 65001 (seg0 slice2 172.31.43.226:7000 pid=14069)

WARNING: bogus varno: 65001 (seg0 slice2 172.31.43.226:7000 pid=14069)

WARNING: bogus varno: 65001 (seg0 slice2 172.31.43.226:7000 pid=14069)

WARNING: bogus varno: 65001 (seg0 slice2 172.31.43.226:7000 pid=14069)

WARNING: bogus varno: 65001 (seg0 slice2 172.31.43.226:7000 pid=14069)

WARNING: bogus varno: 65001 (seg1 slice2 172.31.43.226:7001 pid=14070)

WARNING: bogus varno: 65001 (seg1 slice2 172.31.43.226:7001 pid=14070)

WARNING: bogus varno: 65001 (seg1 slice2 172.31.43.226:7001 pid=14070)

WARNING: bogus varno: 65001 (seg1 slice2 172.31.43.226:7001 pid=14070)

WARNING: bogus varno: 65001 (seg1 slice2 172.31.43.226:7001 pid=14070)

WARNING: bogus varno: 65001 (seg1 slice2 172.31.43.226:7001 pid=14070)

WARNING: bogus varno: 65001 (seg1 slice2 172.31.43.226:7001 pid=14070)

WARNING: bogus varno: 65001 (seg1 slice2 172.31.43.226:7001 pid=14070)

server closed the connection unexpectedly

    This probably means the server terminated abnormally

    before or while processing the request.

The connection to the server was lost. Attempting reset: Failed.

!>

 

I'm attaching the master and segment logs for the above query crash with this ticket.

 

Attachment  View in library