Getting java.sql.SQLException: Error writing file '/tmp/MYNYq5lu' (Errcode: 28)

aashil · 03-03-2010, 10:02 AM

Hello Everyone,

I am working on x86_64-redhat-linux-gnu server and I installed Mysql mysql-standard-5.0.18-linux-i686-glibc23.[I did not installed RPM as I want to be Mysql on my user not as root.]And I am using apache-tomcat-6.0.13.
When I am generating some reports I am getting below error.Also I am unable to find the file where I can mention /tmp.

java.sql.SQLException: Error writing file '/tmp/MYNYq5lu' (Errcode: 28)
at com.mysql.jdbc.MysqlIO.checkErrorPacket(MysqlIO.java:2928)
at com.mysql.jdbc.MysqlIO.sendCommand(MysqlIO.java:1571)
at com.mysql.jdbc.MysqlIO.sqlQueryDirect(MysqlIO.java:1666)
at com.mysql.jdbc.Connection.execSQL(Connection.java:2988)
at com.mysql.jdbc.Connection.execSQL(Connection.java:2917)
at com.mysql.jdbc.Statement.executeQuery(Statement.java:824)
at org.apache.commons.dbcp.DelegatingStatement.executeQuery(DelegatingStatement.java:205)
at com.ft.middleware.rtls.server.db.MARSDataAccess.getServiceData(MARSDataAccess.java:1228)
at com.ft.middleware.rtls.reports.MARSReportGenerator.generateWeeklyReportData(MARSReportGenerator.java :964)
at com.ft.middleware.rtls.reports.MARSReportGenerator.generateWeeklyReport(MARSReportGenerator.java:481 )
at com.ft.middleware.rtls.reports.GenerateWeeklyReports.main(GenerateWeeklyReports.java:29)
Completed DB Query Execution
Total Time Taken For Weekly Report Query [secs] : 144
java.lang.NullPointerException
at com.ft.middleware.rtls.reports.MARSReportGenerator.generateWeeklyReportData(MARSReportGenerator.java :971)
at com.ft.middleware.rtls.reports.MARSReportGenerator.generateWeeklyReport(MARSReportGenerator.java:481 )
at com.ft.middleware.rtls.reports.GenerateWeeklyReports.main(GenerateWeeklyReports.java:29)

Thanks in advance.
Aashil

TB0ne · 03-04-2010, 11:41 AM

Quote:

Originally Posted by aashil

Hello Everyone,

I am working on x86_64-redhat-linux-gnu server and I installed Mysql mysql-standard-5.0.18-linux-i686-glibc23.[I did not installed RPM as I want to be Mysql on my user not as root.]And I am using apache-tomcat-6.0.13.
When I am generating some reports I am getting below error.Also I am unable to find the file where I can mention /tmp.

java.sql.SQLException: Error writing file '/tmp/MYNYq5lu' (Errcode: 28)

Thanks in advance.
Aashil

How about telling us what you're generating reports WITH? What code? Version/distro of Linux?

Since you did a non-standard installation of MySQL, that's probably the cause of your problem. You don't say how you installed it, but assuming you did it from source. Also, the Tomcat engine runs Java code...but who wrote that code? Check it too, but having MySQL run as your own, personal user is a bad idea.

paulsm4 · 11-15-2010, 05:43 PM

Don't mean to necromance here ... but I think the OP's /tmp partition might have been too small.

cjfester · 02-02-2011, 11:40 AM

i have same issue, a table with 145559996 "lines" and 4 fields for group, indexed, and more 4 fields for sum/count. The total table size, in disk, is

Code:

-rw-rw----  1 mysql mysql 8,7K Fev  1 19:38 tmp_ren201009.frm
-rw-rw----  1 mysql mysql 4,5G Fev  1 20:38 tmp_ren201009.MYD
-rw-rw----  1 mysql mysql 2,3G Fev  1 20:38 tmp_ren201009.MYI

uname -r: 2.6.18-194.3.1.el5.centos.plus

mysql --version: mysql Ver 14.14 Distrib 5.1.48, for unknown-linux-gnu (x86_64) using readline 5.1

free -m:

Code:

[root@Controladoria1 ~]# free -m
             total       used       free     shared    buffers     cached
Mem:         48299      35559      12740          0        323      34100
-/+ buffers/cache:       1135      47164
Swap:        24575          0      24575

df -h:

Code:

File System            Size Used Avail Use% Mount at
/dev/sda1              30G  6,6G   22G  24% /          'here is /tmp included
/dev/sdb1             2,3T  653G  1,7T  29% /home      'other disk punch mounted, raid 0 of 6 fisical disks 
/dev/sda3             412G  105G  308G  26% /X         'interchange área with other systems/plataforms
/dev/sdc1             917G  594G  278G  69% /backup    'a external disk atached in usb for backup critical data
tmpfs                  24G     0   24G   0% /dev/shm   'the swap filesystem

i noted the error occurs when system try to flush memory in swap área or /tmp, im not sure, but i see when the memory reaches the "fisical" limit of 48Gb and produces the error describe above.

how can I join the two areas, swap of the system (24Gb) and the /tmp (about 20Gb), in one with ~44Gb ? If /tmp size not enough, the system should not use the swap área ?

A system with 48Gb of memory, ~20Gb disk space for /tmp ang more 24Gb of swap is not enough for rum a "simple" query in a big, not huge, database? is the sky the limit?