/salxlive/base contains seven .lzm files. If I delete 03-full.lzm, will I get "Salix Live basic"? That is forty percent or so of the total size, so I am thinking it might speed up boot (unpacking) time considerably. I could then just add the apps I actually use when on the road, nearly all of which could be extremely lightweight, do an update, and run liveclone to produce a USB with a nippier startup. This is a bit crude, but should it work? Would you need to renumber the remaining .lzm files to be in sequence or anything like that?
The other way round - delete everything you don't use, then use LiveClone - apart from being laborious, would risk missing some or deleting others that are actually needed.
trimming Salix Live
Re: trimming Salix Live
Yes it should work. No need to renumber.
Of course the Salix Installer will not work anymore as you will not be able to install a full Salix from such a modified Live version, but I assume it's not your concern.
It's a good idea if you want to start personalize from "basic" instead of "full". But be aware that Basic...is very basic.
Yes deleting what you don't use could be a bit risky for some packages or you know very well what each package do.
Of course the Salix Installer will not work anymore as you will not be able to install a full Salix from such a modified Live version, but I assume it's not your concern.
It's a good idea if you want to start personalize from "basic" instead of "full". But be aware that Basic...is very basic.
Yes deleting what you don't use could be a bit risky for some packages or you know very well what each package do.
Re: trimming Salix Live
Thanks JRD; I'll give it a go when I have time to fiddle. I also thought of doing a "CoreUSB" for CLI only! 
The one heavy app I might want to include is OpenOffice (just for Writer, but I expect you need to have the whole suite). Am I right in thinking that to achieve a faster boot, the best approach would be to create my USB without OO using LiveClone, then add it using persistence?
The one heavy app I might want to include is OpenOffice (just for Writer, but I expect you need to have the whole suite). Am I right in thinking that to achieve a faster boot, the best approach would be to create my USB without OO using LiveClone, then add it using persistence?
Re: trimming Salix Live
Yes should be quicker as a compressed squashfs4 module will take more time to read than a XFS file. This is so true for OpenOffice as the package is more than 100MB (buffer hunks are 1MB on compressed squashfs4 modules)
Re: trimming Salix Live - success!
Just to confirm that it worked, and booting is now almost reasonable in speed. I'm using Salix LXDE, which must help with that too.
I was slightly taken aback at the absence of wicd - as you say JRD, it's no frills - but discovered how to configure the network manually at www.linuxquestions.org. The Slackware website tells you about netconfig (choose DHCP), but doesn't mention the (no doubt obvious) crucial next step:
(Note the second character in line two is a space.)
Just in case anyone who wants to try this at home also didn't know that
The next thing I did was install wicd (after updating), though maybe I don't really need it any more.
I was slightly taken aback at the absence of wicd - as you say JRD, it's no frills - but discovered how to configure the network manually at www.linuxquestions.org. The Slackware website tells you about netconfig (choose DHCP), but doesn't mention the (no doubt obvious) crucial next step:
Code: Select all
cd /etc/rc.d
. rc.inet1Just in case anyone who wants to try this at home also didn't know that
The next thing I did was install wicd (after updating), though maybe I don't really need it any more.
Re: trimming Salix Live - success!
Cool !mimosa wrote:Just to confirm that it worked, and booting is now almost reasonable in speed. I'm using Salix LXDE, which must help with that too.
Well, in fact you can just domimosa wrote:Code: Select all
cd /etc/rc.d . rc.inet1
Code: Select all
# service restart inet1Re: trimming Salix Live
It also works fine with just core. The only trouble with my scheme is that I suppose the persistence tool (and indeed LiveClone) depends on X - or does it? I left the remaining .lzm fies alone, so the tools must be there. I got as far as installing xfce but don't know what else is needed to actually start the GUI ("startx", no dice). I'll keep fiddling.
Of course I'm just having a bit of fun here, but the serious point is that the imperative for a live CD is data compression to pack as much into ~700MB as you can; once you put the data on a USB key, these days space is less critical, so there is scope to increase boot speed by minimizing compression. In concrete terms, if space is unimportant, it is better for speed to put as much software as possible in uncompressed form. Given the Salix live tools, the obvious way to do so is by creating a bigger persistence file and putting the software there, as I have tentatively been doing. I'm thinking about Salix on USB as a practical working tool when on the road or for new users experimenting. The most salient thing wrong with it now is it takes perhaps five minutes or more to achieve liftoff, at least on slowish hardware.
On the other hand, this has nothing essentially to do with persistence. I can imagine a future version of LiveClone which, for USBs or even DVDs, used less data compression in the interest of faster boot time, while still offering an entirely independent persistence capability for user data and personalised applications.
Of course I'm just having a bit of fun here, but the serious point is that the imperative for a live CD is data compression to pack as much into ~700MB as you can; once you put the data on a USB key, these days space is less critical, so there is scope to increase boot speed by minimizing compression. In concrete terms, if space is unimportant, it is better for speed to put as much software as possible in uncompressed form. Given the Salix live tools, the obvious way to do so is by creating a bigger persistence file and putting the software there, as I have tentatively been doing. I'm thinking about Salix on USB as a practical working tool when on the road or for new users experimenting. The most salient thing wrong with it now is it takes perhaps five minutes or more to achieve liftoff, at least on slowish hardware.
On the other hand, this has nothing essentially to do with persistence. I can imagine a future version of LiveClone which, for USBs or even DVDs, used less data compression in the interest of faster boot time, while still offering an entirely independent persistence capability for user data and personalised applications.