Note: This article assumes you have command-line access to the ESXi host (TechSupport Mode).
To analyze this issue and gather process information:
- Run the minimalized ps command:
ps -Tcjstv
The output shows information such PID, command-line used, and state.
- To generate output for certain processes and maintain a title row, run this command:
ps -Tcjstv | egrep "WID|{proc-name}"
- The esxcfg-info command can offer some detail on process memory and CPU and is contained in the vm-support dumps. Process information in esxcfg-info is contained in sections called Groups, but the groups can be cumbersome to sift through.
Grep these groups for information on process memory and CPU with the command:
esxcfg-info | egrep -A85 "Group Id" | egrep "Group Name|Group Id|Total CPU|Total Memory|Used Time|Parent Id|----Active\.\.\.\." | less
- To check for file-system issues, run this command:
stat –f {filesystem}
Note: This checks inode use, from which a runaway process can sometimes be determined (even if you do not know exactly what files are involved).
To check across all filesystems, run this command:
stat -f `df | grep -v "Filesystem" | awk '{print $6}'` | egrep "File|Inodes"
- To view network connection (socket) details, run this command:
esxcli network connection list
Note: For ESXi 5.x and later, run this command:
esxcli network ip connection list
Note: At the time of publication, there is no way to tell how many open files a process holds (for example, lsof).