Skip to main content

ERROR: ERROR 3587: Insufficient resources to execute plan on pool general

  • October 8, 2015
  • 6 replies
  • 21 views

AxelK

Just downloaded "HP Vertica Analytic Database Server" (to be exact - vertica-7.1.2-0.x86_64.RHEL5). Played a little. M-m-m, strange. Trying to understand Vertica's logic of contraints handling, faced with problem:

Simple sequence causes an issue reported in subject:

 

create table Employees (EmployeeID int primary key, FirstName varchar(30), LastName varchar(30), SSN varchar(30) unique);
create table HealthPlans (HealthPlanID int primary key, EmployeeID int NOT NULL references Employees (EmployeeID), name varchar(30));
CREATE PROJECTION EMP_HEALTH (EmployeeID, FirstName, LastName, SSN)
AS (SELECT e.EmployeeID, e.FirstName, e.LastName, e.SSN
FROM Employees e, HealthPlans hp
WHERE e.EmployeeID = hp.EmployeeID);

 

insert into Employees (EmployeeID, FirstName, LastName, SSN) values (1, 'John', 'Doe', '1111');
insert into Employees (EmployeeID, FirstName, LastName, SSN) values (3, 'Richard', 'Roe', '3333');
commit;
insert into HealthPlans (HealthPlanID, EmployeeID) values (1, 1);
insert into HealthPlans (HealthPlanID, EmployeeID) values (3, 3);
commit;

select * from EMP_HEALTH;

/*As expectable, 2 rows*/

delete from HealthPlans where EmployeeID = 1;

/*delete from child table - OK*/

delete from Employees where EmployeeID = 1;

/*

ERROR: ERROR 3587: Insufficient resources to execute plan on pool
general [Request Too Large:Memory(KB) Exceeded: Requested =
594635, Free = 512827 (Limit = 512827, Used = 0)] Error Code:
3587
Query = delete from Employees where EmployeeID = 1

*/

 

All settings are from default installation (VM was downloaded). Default settings prevent me to add 4 and delete 1-2 records??? 

 

Thanks in advance,

  Axel

6 replies

SruthiA
Forum|alt.badge.img+1
  • Participating Frequently
  • October 14, 2015

Hi,

 

   Can you check  settings for the following parameters in  your nodes

 

/sys/kernel/mm/redhat_transparent_hugepage/defrag

/sys/kernel/mm/redhat_transparent_hugepage/khugepaged/defrag

 

You can find detailed information about those settings in the following URL

 

http://my.vertica.com/docs/7.1.x/HTML/index.htm#Authoring/InstallationGuide/BeforeYouInstall/transparenthugepages.htm?Highlight=hugepage

 

 

Thanks,

 Sruthi


AxelK
  • Author
  • New Participant
  • October 14, 2015

[dbadmin@localhost ~]$ ls /sys/kernel/mm/transparent_hugepage/                  
defrag  enabled  khugepaged  use_zero_page
[dbadmin@localhost ~]$ cat /sys/kernel/mm/transparent_hugepage/enabled          
always madvise [never]
[dbadmin@localhost ~]$ cat /sys/kernel/mm/transparent_hugepage/defrag
always madvise [never]
[dbadmin@localhost ~]$ cat /sys/kernel/mm/transparent_hugepage/khugepaged/defrag

1

 

Seems ok.

 

Thank you,

  Axel


SruthiA
Forum|alt.badge.img+1
  • Participating Frequently
  • October 14, 2015

Hi,

 

 

    Can you share me the output of following queries

 

1. select * from host_resources;

2. select * from resource_pools;

3. select * from resource_pool_status;

 

 

Sruthi


AxelK
  • Author
  • New Participant
  • October 14, 2015

select * from host_resources;

 

host_name    open_files_limit    threads_limit    core_file_limit_max_size_bytes    processor_count    processor_core_count    processor_description    opened_file_count    opened_socket_count    opened_nonfile_nonsocket_count    total_memory_bytes    total_memory_free_bytes    total_buffer_memory_bytes    total_memory_cache_bytes    total_swap_memory_bytes    total_swap_memory_free_bytes    disk_space_free_mb    disk_space_used_mb    disk_space_total_mb
127.0.0.1 65536 4096 0 2 2 Common KVM processor 7 11 6 1042182144 68186112 50089984 437104640 2164256768 1991012352 40330 7412 47743

select * from resource_pools;

 

pool_id              name        is_internal    memorysize    maxmemorysize    executionparallelism    priority    runtimepriority    runtimeprioritythreshold    queuetimeout    plannedconcurrency    maxconcurrency    runtimecap    singleinitiator    cpuaffinityset    cpuaffinitymode    cascadeto
45035996273720528 general 1 Special: 95% AUTO 0 MEDIUM 2 300 AUTO 0 ANY
45035996273720530 sysquery 1 64M AUTO 110 HIGH 0 300 AUTO 0 ANY
45035996273720532 sysdata 1 100M 10%
45035996273720534 wosdata 1 0% 25% AUTO
45035996273720536 tm 1 200M AUTO 105 MEDIUM 60 300 AUTO 3 1 ANY
45035996273720538 refresh 1 0% AUTO -10 MEDIUM 60 300 AUTO 1 ANY
45035996273720540 recovery 1 0% AUTO 107 MEDIUM 60 300 AUTO 1 1 ANY
45035996273720542 dbd 1 0% AUTO 0 MEDIUM 0 0 AUTO 1 ANY
45035996273720624 jvm 1 0% 10% AUTO 0 MEDIUM 2 300 AUTO 0 ANY

select * from resource_pool_status;

 

node_name          pool_oid             pool_name    is_internal    memory_size_kb    memory_size_actual_kb    memory_inuse_kb    general_memory_borrowed_kb    queueing_threshold_kb    max_memory_size_kb    running_query_count    planned_concurrency    max_concurrency    is_standalone    queue_timeout_in_seconds    execution_parallelism    priority    runtime_priority    runtime_priority_threshold    single_initiator    query_budget_kb    cpu_affinity_set    cpu_affinity_mask    cpu_affinity_mode
v_test_node0001 45035996273720528 general 1 512827 512827 0 0 487185 512827 0 4 1 300 AUTO 0 MEDIUM 2 false 121796 3 ANY
v_test_node0001 45035996273720530 sysquery 1 65536 65536 0 0 549444 578363 0 4 0 300 AUTO 110 HIGH 0 false 16384 3 ANY
v_test_node0001 45035996273720532 sysdata 1 87018 87018 8192 0 82667 87018 0 1 0 1 0 AUTO 0 HIGH 0 false 3 ANY
v_test_node0001 45035996273720534 wosdata 1 0 0 0 0 206667 217545 0 2 0 0 0 AUTO 0 HIGH 0 false 3 ANY
v_test_node0001 45035996273720536 tm 1 204800 204800 0 0 681745 717627 0 2 3 0 300 AUTO 105 MEDIUM 60 true 102400 3 ANY
v_test_node0001 45035996273720538 refresh 1 0 0 0 0 487185 512827 0 4 0 300 AUTO -10 MEDIUM 60 true 121796 3 ANY
v_test_node0001 45035996273720540 recovery 1 0 0 0 0 487185 512827 0 4 1 0 300 AUTO 107 MEDIUM 60 true 121796 3 ANY
v_test_node0001 45035996273720542 dbd 1 0 0 0 0 487185 512827 0 4 0 0 AUTO 0 MEDIUM 0 true 121796 3 ANY
v_test_node0001 45035996273720624 jvm 1 0 0 0 0 82667 87018 0 4 0 300 AUTO 0 MEDIUM 2 false 20666 3 ANY

Again, all of parameters left unchanged - as they were configured in originally downloaded VM.

 

Thanks,

  Axel

 


SruthiA
Forum|alt.badge.img+1
  • Participating Frequently
  • October 14, 2015

Hi,

 

  Vertica recommends a minimum of 4GB of memory per core. i see that you have 2 cores and just 1 GB of memory. Memory allocation is very low. You need to have 2*4B = 8GB of memory. I request you to provided atleast 8GB of memory  to your cluster to solve this issue

 

 

Sruthi

 

 


AxelK
  • Author
  • New Participant
  • October 15, 2015

SruthiA, have you read my messages?

 

"All settings are from default installation (VM was downloaded). Default settings prevent me to add 4 and delete 1-2 records??? "

 

"Again, all of parameters left unchanged - as they were configured in originally downloaded VM."

 

I'm talking about preconfigured VM which is distributed to clients. For sure I have increased already RAM size to continue my checks on another VM instance - but still, the question remains: is it OK when failure appears on just two-table-four-rows db with 512827 KB (!) of memory available for query processing? That affects on level of confidence in engine drastically...

 

Bye, Axel