Posted on September 16, 2020 12:32 am
 |  Asked by Patsadon Sangsuwan
 |  35 views
RESOLVED
0
0
Print Friendly, PDF & Email

Hi

I’m curious when I hit show log and Agent log comes out as” warm start”.
Anyone know what does it means?
Also If anyone can clarify how an Agent state works, I would be really appreciated

Thank you in advance

0
Posted by Alexis Dacquay
Answered on September 16, 2020 2:32 pm

Hi Patsadon,
Can you share the complete log message?
Typically it's for an agent restarting because you configured something that changes the agent's behaviour.

You can see the agent status with "show agent <...>"

Does this respond to the question, or are you after how Agents in EOS work? It's a long and nice story, but I recommend you contact your SE or try to find one a Systems Engineer via sales@arista.com to start the conversation.

Regards,
Alexis

0
Posted by Alexis Dacquay
Answered on September 16, 2020 2:33 pm

Here to start with, about EOS agents:
https://www.arista.com/assets/data/pdf/Whitepapers/Arista_Advantage.pdf

0
Posted by Vignesh
Answered on September 16, 2020 6:23 pm

Hi Patsadon,

Thanks for writing to Arista.

ProcMgr is a python script .On startup, ProcMgr reads its configuration file "/etc/ProcMgr.conf" and ProcMgr forks,
* The parent process (the "ProcMgr (master)") sits in a very simple loop. This is the process that restarts ProcMgr if it crashes.
* The child process (the "ProcMgr (worker)") is the ProcMgr worker process that does the heavy lifting of spawning configured processes, setting up their environment variables, monitoring their health, etc. etc.If this process crashes, the bereaved master process forks another worker, which takes over where the previous worker left off.
 ProcMgr worker performs a warm start on startup for the processes ( agents) managed by it  .This can be expected to happen when the switch initially boots up or if any agent managed by the procmgr worker is restarted or started newly due to turning on a feature .The Procmgr worker monitors heartbeats from the various processes controller by it . When ProcMgr receives SIGHUP, it performs a warm restart as described above. This covers the case of upgrading the system, or adding or removing services

For example ,

This is when the switch booted for the first time

 

Fri Sep 04 02:39:05.224728 2020: PROCMGR_PROCESS_STARTED: 'VxlanSwFwd' starting with PID=5163 (PPID=1887) -- execing '/usr/bin/VxlanSwFwd'
Fri Sep 04 02:39:05.225375 2020: PROCMGR_WORKER_WARMSTART_DONE: ProcMgr worker warm start done. (PID=1887)

 

This is when a new feature is configured

Sun Sep 13 23:22:51.989010 2020: PROCMGR_PROCESS_STARTED: 'Sflow' starting with PID=17430 (PPID=1887) -- execing '/usr/bin/Sflow'
Sun Sep 13 23:22:51.989588 2020: PROCMGR_WORKER_WARMSTART_DONE: ProcMgr worker warm start done. (PID=1887)

I hope this helps.

Thanks ,

Vignesh Jothinarayanan

Post your Answer

You must be logged in to post an answer.