Linux Server JPS report process information unavailable

Execution of JPS under Linux is a quick view of the Java program process commands, generally hadoop,hbase,storm and other processes are viewed through JPS, sometimes because the process does not end normally, For example, when the resource is too large to hang off or not to end the process to restart the computer directly, it will cause the original process to become--process information unavailable such a null value, sometimes do not control it, a period of time will automatically disappear, if not disappear in the case, You can use the following methods to clean up:

Enter/tmp directory cd/tmp can see a lot of hsperfdata_{user name} Such directory, such as: Hsperfdata_hbase,hsperfdata_kafka,hsperfdata_root such directory , because the process is closed in memory, but Linux will also look for these temporary files in/tmp, while the temporary files are not deleted properly, this time directly to execute rm -rf hsperfdata_* Delete these directories, Then perform the JPS view again, and those processes will not exist.

In general, direct execution of rm -rf/tmp/hsperfdata_* can quickly remove those residue processes

If there are a lot of normal running processes, which are part of the residual process, it is not recommended to delete all of the above directory, this time, according to the directory and process of the corresponding relationship analysis of the remaining single directory to remove the useless process can be

Linux Server JPS report process information unavailable