RE: vmstat -s "boot time" Redhat Linux

From: Matthew Zito <mzito_at_gridapp.com>
Date: Fri, 13 Nov 2009 12:05:17 -0500
Message-ID: <C0A5E31718FC064A91E9FD7BE2F081B101D4BC5E_at_exchange.gridapp.com>


Well the boot time really shouldn't update if the machine hasn't rebooted in a physical environment - its not the duration since the boot, it's the specific boot time. In a virtual environment, you could be right, though theoretically in a non-paravirtualized environment it shouldn't be aware of moving to another host. Just a sanity check - the uptime doesn't change as part of that, does it?

Matt

-----Original Message-----
From: Ethan Post [mailto:post.ethan_at_gmail.com] Sent: Fri 11/13/2009 12:02 PM
To: Matthew Zito
Cc: oracle-l
Subject: Re: vmstat -s "boot time" Redhat Linux  

Maybe this is more VMWARE weirdness. The number is static for long periods of time, and then after a few days goes up (perhaps when the image is migrated to another server??).

$ while ((1)); do
> vmstat -s | grep "boot time"
> sleep 600
> done

   1254804111 boot time
   1254804111 boot time

On Fri, Nov 13, 2009 at 10:05 AM, Matthew Zito <mzito_at_gridapp.com> wrote:

>
> That is the actual time in seconds since the epoch (unix time) that the
> machine was started. It's stored in /proc/stat, and vmstat just grabs the
> raw value and displays it with a slightly better label (it's btime in
> /proc/stat).
>
> So, in other words, take that unix time, convert it to a human-readable
> time, adn you've got when the machine was booted.
>
> Thanks,
> Matt
>
> --
> Matthew Zito
> Chief Scientist
> GridApp Systems
> P: 646-452-4090
> mzito_at_gridapp.com
> http://www.gridapp.com
>
>

--
http://www.freelists.org/webpage/oracle-l
Received on Fri Nov 13 2009 - 11:05:17 CST

Original text of this message