[MCOL-689] cpimport error Created: 2017-04-28  Updated: 2017-05-02  Resolved: 2017-05-02

Status: Closed
Project: MariaDB ColumnStore
Component/s: writeengine
Affects Version/s: 1.0.8
Fix Version/s: Icebox

Type: Bug Priority: Major
Reporter: Seungmin Yang Assignee: Unassigned
Resolution: Not a Bug Votes: 0
Labels: None
Environment:

RHEL 6.6

Attachments: File columnstoreSupportReport.columnstore-1.tar.gz    

 Description   

Hi, I had an error while using cpimport command. I configured separate modules with 2 PMs and 1 UM. The type of data storage is external. I executed cpimport command on PM #1(parent OAM). Load file(test_1) contains 10,000,000 rows data.

Error messages are belows:

[root@cm1 ~]# cpimport input_test mf2ae_log test_1
2017-03-31 02:58:17 (11610) INFO : Running distributed import (mode 1) on all PMs...
2017-03-31 02:59:07 (11610) ERR : Received a Cpimport Failure from PM1
2017-03-31 02:59:07 (11610) INFO : Please verify error log files in PM1
2017-03-31 02:59:07 (11610) INFO : Canceling outstanding cpimports
2017-03-31 03:00:12 (11610) INFO : Table input_test.mf2ae_log: (OID-3000) was NOT successfully loaded.
2017-03-31 03:00:12 (11610) INFO : Bulk load completed, total run time : 115.004 seconds

Log messages(err.log, debug.log):

[root@cm1 ~]# cat /var/log/mariadb/columnstore/err.log
Mar 31 02:59:04 cm1 writeengineserver[3499]: 04.404181 |0|0|0| E 32 CAL0000: pushing data : PIPE error .........Broken pipe
Mar 31 02:59:07 cm1 writeengineserver[3499]: 07.440615 |0|0|0| E 32 CAL0000: 1905 : cpimport exit on failure (signal 9)
Mar 31 02:59:07 cm1 writeenginesplit[11610]: 07.440784 |0|0|0| E 33 CAL0000: #033[0;31mReceived a Cpimport Failure from PM1#033[0m
Mar 31 02:59:07 cm1 writeenginesplit[11610]: 07.491808 |0|0|0| E 33 CAL0087: BulkLoad Error: #033[0;31mReceived a Cpimport Failure from PM1#033[0m

/var/log/mariadb/columnstore/debug.log

Mar 31 02:58:17 cm1 writeengineserver[3499]: 17.153209 |0|0|0| D 32 CAL0000: 1905 : onReceiveMode() Setting fMode = 1
Mar 31 02:58:17 cm1 writeengineserver[3499]: 17.154965 |0|0|0| D 32 CAL0000: 1905 : onReceiveMode() DbRoot Count = 1
Mar 31 02:58:17 cm1 writeengineserver[3499]: 17.155080 |0|0|0| D 32 CAL0000: 1905 : CMD LINE ARGS came in /usr/local/mariadb/columnstore/bin/cpimport.bin -R /tmp/columnstore_tmp_files/BrmRpt023102581711610.rpt -m 1 -P pm1-11610 -u6a3eae55-54ee-4b59-9114-512e37f363bf input_test mf2ae_log
Mar 31 02:58:17 cm1 writeengineserver[3499]: 17.157575 |0|0|0| D 32 CAL0000: 1905 : Brm Rpt Filename Arrived /tmp/columnstore_tmp_files/BrmRpt023102581711610.rpt
Mar 31 02:58:17 cm1 writeengineserver[3499]: 17.157896 |0|0|0| D 32 CAL0000: 1905 : Start Cpimport command reached!!
Mar 31 02:59:07 cm1 writeengineserver[3499]: 07.324109 |0|0|0| D 32 CAL0000: 1905 : onReceiveEOD : child ID = 11644
Mar 31 02:59:07 cm1 writeengineserver[3499]: 07.324183 |0|0|0| D 32 CAL0000: 1905 : Message Queue is empty; Stopping CF Thread
Mar 31 02:59:07 cm1 writeengineserver[3499]: 07.465547 |0|0|0| D 32 CAL0000: 1905 : onCpimportFailure BrmReport Send
Mar 31 02:59:07 cm1 writeengineserver[3499]: 07.465674 |0|0|0| D 32 CAL0000: 1905 : onReceiveEOD : child ID = 0
Mar 31 02:59:07 cm1 writeengineserver[3499]: 07.465759 |0|0|0| D 32 CAL0000: 1905 : onReceiveEOD : child ID = 0
Mar 31 03:00:09 cm1 writeengineserver[3499]: 09.655294 |0|0|0| D 32 CAL0000: 1905 : OnReceiveRollback arrived
Mar 31 03:00:11 cm1 writeengineserver[3499]: 11.998767 |0|0|0| D 32 CAL0000: 1905 : OnReceiveCleanup arrived

Could you let me know why this error occured?

 Comments   
Comment by Andrew Hutchings (Inactive) [ 2017-04-28 ]

Can you please run the MariaDB ColumnStore support tool and attach the resulting .tar.gz? This will help give us a clearer picture of what is going on across the cluster at the time of failure:

https://mariadb.com/kb/en/mariadb/system-troubleshooting-mariadb-columnstore/#mariadb-columnstore-support-tool

Comment by Seungmin Yang [ 2017-05-02 ]

I attached ColumnStore support report file.
Thank you very much!

Comment by David Thompson (Inactive) [ 2017-05-02 ]

Hi, in looking at your hardware config you are running 1g memory 1 core servers for PMs. While this can work for simple demos it can run into trouble with larger data sets. Normally we'd recommend an absolute minimum of 2gb and more like 4gb per server for a dev environment. I believe what happened is one of the servers went into heavy swap and so became non responsive breaking the remote connection (broken pipe error) causing the cpimport to fail.

Generated at Thu Feb 08 02:23:05 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.