Date: 2013-09-07 06:41:59 +0200
From: Tim H. <<monetdb.bug.reporter>>
To: SQL devs <>
Version: 11.15.7 (Feb2013-SP2)
CC: @njnes
Last updated: 2013-09-27 13:47:14 +0200
Comment 19139
Date: 2013-09-07 06:41:59 +0200
From: Tim H. <<monetdb.bug.reporter>>
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/29.0.1547.66 Safari/537.36
Build Identifier:
Start mserver5 so that it initializes a new database. Then load the TPCH schema plus keys. Finally, plan the TPCH queries. Do this twice, each time with a newly initialized database. Diffing the two outputs will show where uninitialized memory is written as a string in the plan.
568,570c568,570
< | | | | | table(sys.customer) [ customer.c_custkey NOT NULL HASHCOL as �.c_custkey, customer.c_name NOT NULL as �.c_name, customer.%TID% NOT NULL as �.%TID%, customer.%customer_c_custkey_pkey NOT NULL HASHIDX as �.%customer_c_custkey_pkey, customer.%customer_c_nationkey_fkey NOT NULL JOINIDX sys.customer.customer_c_nationkey_fkey as �.%customer_c_nationkey_fkey ] COUNT
< | | | | ) [ customer.c_custkey NOT NULL HASHCOL = �.c_custkey NOT NULL HASHCOL FETCH ]
< | | | ) [ customer.c_custkey NOT NULL HASHCOL , orders.o_orderkey NOT NULL HASHCOL , L6.L6 NOT NULL, �.c_custkey NOT NULL, �.c_name NOT NULL as customer.c_name ],
| | | | | table(sys.customer) [ customer.c_custkey NOT NULL HASHCOL as �.c_custkey, customer.c_name NOT NULL as �.c_name, customer.%TID% NOT NULL as �.%TID%, customer.%customer_c_custkey_pkey NOT NULL HASHIDX as �.%customer_c_custkey_pkey, customer.%customer_c_nationkey_fkey NOT NULL JOINIDX sys.customer.customer_c_nationkey_fkey as �.%customer_c_nationkey_fkey ] COUNT
| | | | ) [ customer.c_custkey NOT NULL HASHCOL = �.c_custkey NOT NULL HASHCOL FETCH ]
| | | ) [ customer.c_custkey NOT NULL HASHCOL , orders.o_orderkey NOT NULL HASHCOL , L6.L6 NOT NULL, �.c_custkey NOT NULL, �.c_name NOT NULL as customer.c_name ],
Expected Results:
This is with a build of the MonetDB repository Feb2013 branch @48758, configured with all the defaults. I.e. just ./configure. Note that there is a SQL file referenced in the repro steps. I will attach it to this bug.
I'm marking this as critical because memory issues in C programs can cause bad things to happen.
Comment 19140
Date: 2013-09-07 06:42:54 +0200
From: Tim H. <<monetdb.bug.reporter>>
Created attachment 232
SQL file referenced in the repro steps
Attached file: tpch_query_plans.sql (application/octet-stream, 13914 bytes)
Description: SQL file referenced in the repro steps
Date: 2013-09-07 06:41:59 +0200
From: Tim H. <<monetdb.bug.reporter>>
To: SQL devs <>
Version: 11.15.7 (Feb2013-SP2)
CC: @njnes
Last updated: 2013-09-27 13:47:14 +0200
Comment 19139
Date: 2013-09-07 06:41:59 +0200
From: Tim H. <<monetdb.bug.reporter>>
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/29.0.1547.66 Safari/537.36
Build Identifier:
Start mserver5 so that it initializes a new database. Then load the TPCH schema plus keys. Finally, plan the TPCH queries. Do this twice, each time with a newly initialized database. Diffing the two outputs will show where uninitialized memory is written as a string in the plan.
Reproducible: Always
Steps to Reproduce:
mkdir /tmp/mtest-19378/
rm -rf /tmp/db-19378
mserver5 --dbpath=/tmp/db-19378 --debug=10 --set gdk_nr_threads=0 --set mapi_open=true --set mapi_port=35809 --set mapi_usock=/tmp/mtest-19378/.s.monetdb.35809 --set monet_prompt= --forcemito --set mal_listing=2 &
mclient -lsql -ftest -Eutf-8 -i -e --port=35809 < tpch_query_plans.sql > run1
pkill mserver5
rm -rf /tmp/db-19378
mserver5 --dbpath=/tmp/db-19378 --debug=10 --set gdk_nr_threads=0 --set mapi_open=true --set mapi_port=35809 --set mapi_usock=/tmp/mtest-19378/.s.monetdb.35809 --set monet_prompt= --forcemito --set mal_listing=2 &
mclient -lsql -ftest -Eutf-8 -i -e --port=35809 < tpch_query_plans.sql > run2
diff run1 run2
Actual Results:
568,570c568,570
< | | | | | table(sys.customer) [ customer.c_custkey NOT NULL HASHCOL as �.c_custkey, customer.c_name NOT NULL as �.c_name, customer.%TID% NOT NULL as �.%TID%, customer.%customer_c_custkey_pkey NOT NULL HASHIDX as �.%customer_c_custkey_pkey, customer.%customer_c_nationkey_fkey NOT NULL JOINIDX sys.customer.customer_c_nationkey_fkey as �.%customer_c_nationkey_fkey ] COUNT
< | | | | ) [ customer.c_custkey NOT NULL HASHCOL = �.c_custkey NOT NULL HASHCOL FETCH ]
< | | | ) [ customer.c_custkey NOT NULL HASHCOL , orders.o_orderkey NOT NULL HASHCOL , L6.L6 NOT NULL, �.c_custkey NOT NULL, �.c_name NOT NULL as customer.c_name ],
Expected Results:
This is with a build of the MonetDB repository Feb2013 branch @48758, configured with all the defaults. I.e. just ./configure. Note that there is a SQL file referenced in the repro steps. I will attach it to this bug.
I'm marking this as critical because memory issues in C programs can cause bad things to happen.
Comment 19140
Date: 2013-09-07 06:42:54 +0200
From: Tim H. <<monetdb.bug.reporter>>
Created attachment 232
SQL file referenced in the repro steps
Comment 19141
Date: 2013-09-09 02:36:00 +0200
From: @njnes
fixed, added missing strdup
Comment 19142
Date: 2013-09-09 02:36:58 +0200
From: MonetDB Mercurial Repository <>
Changeset 23a797ab1392 made by Niels Nes niels@cwi.nl in the MonetDB repo, refers to this bug.
For complete details, see http//devmonetdborg/hg/MonetDB?cmd=changeset;node=23a797ab1392
Changeset description:
The text was updated successfully, but these errors were encountered: