After upgrading to 24.2R1-S2.5, this code introduced a very chatty script constantly running in the background that will make the interactive commands rollover with high frequency.
e.g.
May 9 14:46:58 switch mgd[43956]: UI_JUNOSCRIPT_CMD: User 'root' used JUNOScript client to run command 'show system processes host-processes'
May 9 14:46:59 switch mgd[43961]: UI_CMDLINE_READ_LINE: User 'root', command 'show system process extensive '
May 9 14:47:06 switch mgd[43982]: UI_CMDLINE_READ_LINE: User 'root', command 'show bfd session detail '
May 9 14:47:06 switch mgd[43985]: UI_CMDLINE_READ_LINE: User 'root', command 'show ppm adjacencies '
May 9 14:47:06 switch mgd[43988]: UI_CMDLINE_READ_LINE: User 'root', command 'show system processes host-processes '
THE NEED FOR THIS SCRIPT:
At present, in some instances, by the time the customer opens the JTAC case and attaches the RSI and /var/log, the problem is already gone, and we do not have a historic system state at frequent intervals to correlate where the bottleneck could have been. We are collecting the system state via various counters, which will help us assess the potential reason for customer-reported issues.
If we collect key drop counters for protocols like LACP, BFD, and collect the RE/Kernel/PFE stats at regular intervals, we will have more information for us to find why the protocol flapped (due to congestion in the host path or CPU starvation).
These logs and any UI_ log related to the root user are harmless and are safe to be ignored or suppressed e.g.
root@switch# show system syslog
May 09 15:09:35
file interactive-commands {
match "!(.*UI_JUNOSCRIPT_CMD: User 'root'*|.UI_CMDLINE_READ_LINE: User 'root'*)";
}
{master:0}[edit]
root@switch#
May 9 14:47:06 switch mgd[43989]: UI_CMDLINE_READ_LINE: User 'root', command 'command rpc rpc command show system processes host-processes no-forwarding
If, after knowing this, it is still desired to proceed by disabling the script, these are the steps to do so:
From RE shell:
ps -aux |grep system_counters_script.sh
kill -9 <pid>
Note:
1 We must do it from all FPCs because this script runs on each FPC of the VC.
2 The switch will generate a new script after a reboot; hence, you will have to kill it again.
2025-07-10 : Article Created by dcole
2025-08-08 : Article modification.