New 0.5 draft of the SemPiTernal HOWTO

Frank frank at nicholasfamilycentral.com
Fri May 6 14:18:36 UTC 2016


> On May 6, 2016, at 9:45 AM, Eric S. Raymond <esr at thyrsus.com> wrote:
> 
> I'm seeing much faster bootup than this - I don't think I've had to wait
> for ssh availability longer than 30 seconds.  I wonder what accounts
> for the difference?

Sorry - The time I mentioned above was for shutdown & reboot.  The PI I’m working with has both wired & wireless connected.  I intended to make a public NTP server available on wired (separate network) & have admin access via WiFi.  Those times are for the WiFi interface to come up & be available.  I concur that without WiFi, my times are close to your observed times.

SD card model can greatly affect boot time.  A “faster” card my not be quicker for the Pi’s.  The Pi’s have an odd SD card interface back to the SoC & some cards just don’t perform as well as others.

>> Is “Wait for Network at Boot” really required?  On my Gentoo Pi's
>> with dhcpcd, it will background wait indefinitely (maybe my Gentoo
>> configuration setting) , and will grab an IP when a cable is
>> connected.
> 
> I'm not sure if it's necessary, but it seems like a good belt-and-suspenders
> move.

Understood - I just wanted you to know it wasn’t necessarily a requirement.

>> “/boot” is not mounted by default.  To check for overlays, they may
>> have to mount “/boot”
> 
> It's premounted in the lastest Jessie Lite image.  Under what circumstances
> might it not be?

I never mount “/boot” by default (for safety).  I thought that was how Raspian was, but I haven’t checked lately.  I have a card with the full Raspian, and I’ll check this evening.

>> I’m not following the “blink” table.  It’s not jumping out at me or
>> making sense…
> 
> LED blink interval in seconds - telly you fix vs no fix.

Right - *I* understand how the LED works.  What are the values in the table - Blinks per second?  Seconds between blinks?

>> For cloning a GIT repository, I’ve used “- - depth=1” to get what I
>> think is just the top, latest version?  If my understanding is
>> correct, this will be quicker for people on slow connections.  If
>> I’m wrong about what this does in GIT, please correct my
>> understanding.  Is there a down side to using depth=1?
> 
> Will a later git pull work on a repo pulled this way?

I don’t know.  For my purposes, I usually wipe out the previous clone & clone again.  I’m a git novice.  I started using “depth=1” for large repos.

If you tell me how to test, I’m glad to find the answer.

Thanks,
Frank



More information about the devel mailing list