Re: Deprecating NBD in favor of squashfs-over-NFS
Alkis Georgopoulos <[email protected]> Fri, 17 May 2019 23:34:05 +0300
| Newsgroups | gmane.linux.terminal-server.general |
|---|---|
| Message-ID | <[email protected]> |
Hi Bob,
On 5/17/19 11:07 PM, Bob Wicksall wrote:
> One question though. We have a couple wireless bridges in production providing connectivity to a handful of wireless clients. Performance over the bridge takes a pretty big hit so I use "LTSP_NBD_TO_RAM=true" to load the image into RAM. This works well. Will this option go away with NBD? Is there an alternate way to load the image into RAM?
The current code doesn't support LTSP_NBD_TO_RAM from an NFS root,
but it would only take a few lines to implement it.
I.e. instead of copying from /dev/nbd0, it would copy from "$ltsploop".
The code currently lives in
/usr/share/initramfs-tools/scripts/init-bottom/ltsp
> On a related note. My wireless clients take about twice as long to boot using NFS over the bridge. Once the client is booted performance is about the same.
Thank you, these are useful notes, I only benchmarked NFS vs NBD on
gigabit LAN, nowhere else.
_____________________________________________________________________
Ltsp-discuss mailing list. To un-subscribe, or change prefs, goto:
https://lists.sourceforge.net/lists/listinfo/ltsp-discuss
For additional LTSP help, try #ltsp channel on irc.freenode.net