[MAPREDUCE-3583] ProcfsBasedProcessTree#constructProcessInfo() may throw NumberFormatException

• Type: Bug
• Status: Closed
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 0.20.205.0
• Fix Version/s: 1.0.2, 0.23.2

Environment

64-bit Linux:
asf011.sp2.ygridcore.net
Linux asf011.sp2.ygridcore.net 2.6.32-33-server #71-Ubuntu SMP Wed Jul 20 17:42:25 UTC 2011 x86_64 GNU/Linux

• Hadoop Flags: Reviewed
• Assignee: Ted Yu
• Reporter: Ted Yu
• Created: 20/Dec/11 15:07
• Updated: 09/Jan/14 17:22
• Resolved: 23/Feb/12 23:03

Description

HBase PreCommit builds frequently gave us NumberFormatException.

From https://builds.apache.org/job/PreCommit-HBASE-Build/553//testReport/org.apache.hadoop.hbase.mapreduce/TestHFileOutputFormat/testMRIncrementalLoad/:

2011-12-20 01:44:01,180 WARN  [main] mapred.JobClient(784): No job jar file set.  User classes may not be found. See JobConf(Class) or JobConf#setJar(String).
java.lang.NumberFormatException: For input string: "18446743988060683582"
	at java.lang.NumberFormatException.forInputString(NumberFormatException.java:48)
	at java.lang.Long.parseLong(Long.java:422)
	at java.lang.Long.parseLong(Long.java:468)
	at org.apache.hadoop.util.ProcfsBasedProcessTree.constructProcessInfo(ProcfsBasedProcessTree.java:413)
	at org.apache.hadoop.util.ProcfsBasedProcessTree.getProcessTree(ProcfsBasedProcessTree.java:148)
	at org.apache.hadoop.util.LinuxResourceCalculatorPlugin.getProcResourceValues(LinuxResourceCalculatorPlugin.java:401)
	at org.apache.hadoop.mapred.Task.initialize(Task.java:536)
	at org.apache.hadoop.mapred.MapTask.run(MapTask.java:353)
	at org.apache.hadoop.mapred.Child$4.run(Child.java:255)
	at java.security.AccessController.doPrivileged(Native Method)
	at javax.security.auth.Subject.doAs(Subject.java:396)
	at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1083)
	at org.apache.hadoop.mapred.Child.main(Child.java:249)

From hadoop 0.20.205 source code, looks like ppid was 18446743988060683582, causing NFE:

        // Set (name) (ppid) (pgrpId) (session) (utime) (stime) (vsize) (rss)
         pinfo.updateProcessInfo(m.group(2), Integer.parseInt(m.group(3)),

You can find information on the OS at the beginning of https://builds.apache.org/job/PreCommit-HBASE-Build/553/console:

asf011.sp2.ygridcore.net
Linux asf011.sp2.ygridcore.net 2.6.32-33-server #71-Ubuntu SMP Wed Jul 20 17:42:25 UTC 2011 x86_64 GNU/Linux
core file size          (blocks, -c) 0
data seg size           (kbytes, -d) unlimited
scheduling priority             (-e) 20
file size               (blocks, -f) unlimited
pending signals                 (-i) 16382
max locked memory       (kbytes, -l) 64
max memory size         (kbytes, -m) unlimited
open files                      (-n) 60000
pipe size            (512 bytes, -p) 8
POSIX message queues     (bytes, -q) 819200
real-time priority              (-r) 0
stack size              (kbytes, -s) 8192
cpu time               (seconds, -t) unlimited
max user processes              (-u) 2048
virtual memory          (kbytes, -v) unlimited
file locks                      (-x) unlimited
60000
Running in Jenkins mode

From Nicolas Sze:

It looks like that the ppid is a 64-bit positive integer but Java long is signed and so only works with 63-bit positive integers.  In your case,

  2^64 > 18446743988060683582 > 2^63.

Therefore, there is a NFE. 

I propose changing allProcessInfo to Map<String, ProcessInfo> so that we don't encounter this problem by avoiding parsing large integer.

Issue Links

• duplicates: MAPREDUCE-3836 TestContainersMonitor failing intermittently
• is related to: MAPREDUCE-1201 Make ProcfsBasedProcessTree collect CPU usage information
• relates to: MAPREDUCE-5715 ProcfsBasedProcessTree#constructProcessInfo() can still throw NumberFormatException