finally I managed to create and run a peronalized antiX-12 snapshot on the target machine : a thin client with ViaC7 -1000Hz CPU, 256 MB RAM and no disk.
I boot this machine from a 8 GB USB-stick with the iso (~ 290 MB) on the 2.5 GB /dev/sda1 and installed a 2nd partion with ext2 and a 3rd with fat32 (both not in use at the moment) but no swap partition. So, the system should run without swap at the moment, but I never disabled any swap on the system I used for the creation of the snapshot (I did this on a installation of antiX-base in virtualbox).

Actually, I think it is running well -- but I have the feeling, it is MUCH slower than the original (but not customized) Live-USB.
Is there any special"trick" to get the same performance with the snapshot version. Or any tipps for a better (optimized __{{emoticon}}__ ) handlings of the space on the USB-stick ?

Thanks for any suggestion -- I don't have much experience with live systems and thin clients. So, I'm not sure whether this kind of installation makes semnse at all. However, I'm happy to see the machine running, as it is completely fanless ..........

abcdelix wrote:...run a peronalized antiX-12 snapshot on the target machine : a thin client with ViaC7 -1000Hz CPU, 256 MB RAM and no disk.
I boot this machine from a 8 GB USB-stick... but no swap partition.
... I have the feeling, it is MUCH slower than the original (but not customized) Live-USB.

I don't use the snapshot feature so cannot comment directly on it but a couple of ideas do come to mind.

Is the speed comparision being made having booted the thin client with both"original" and the snapshot USB? If only the snapshot version has been used with the thin client it may not be a software only issue. If the thin client has slow (v1.1) USB ports it might easily be seen as slower due to each read and write having to cross the port. Also, different brands of USB flash drive work at different speeds.

256MB RAM may also be a restricting factor if you are trying to run apps which make reasonably high demands for RAM. This can sometimes be seen when loading such an app as the loading process can use more RAM than when the app is running. As you are running from USB (flash drive?) you may not want a disk based swap system, but might benefit from a RAM based swap space. antiX includes zRam for this. Have a look here viewtopic.php?f=4&t=3536&hilit=zram

Is the speed comparision being made having booted the thin client with both"original" and the snapshot USB? If only the snapshot version has been used with the thin client it may not be a software only issue. If the thin client has slow (v1.1) USB ports it might easily be seen as slower due to each read and write having to cross the port. Also, different brands of USB flash drive work at different speeds.

yes, I booted the original iso from the stick and replaced it later by the snapshot. So, the hardware is exactely the same. I noticed the phenomenon when I visited some of the webpages I usually read.

I'm wll aware of the limitations of the system and use the RAM very carefully (i.e.deleted history often; only one app at a time; no cache for the browser and so on)

but might benefit from a RAM based swap space. antiX includes zRam for this.

Thanks for that !!!
I'll give it a try

I followed the instructions to install melodie's zram script and was able to start the script manually. However, I got the message, that my kernel wasn't compiled with the zram option (?????). Is it a matter of the antiX-base.iso or of the virtualbox method ?

Anyway, with the Xtralean cheatcode now everything is much faster now -- guess, that's a really important option on such low-specs machine as mine.....

Out of curiosity, which zram init script are you using? I noticed there are at least three different versions floating around. The one in that forums thread, as well as the one in /usr/local/bin on the core iso wouldn't work for me. They would fail out and complain that the kernel didn't support zram, even though it did. The script in /usr/local/bin on the base iso I found to function properly.

abcdelix wrote:I followed the instructions to install melodie's zram script and was able to start the script manually. However, I got the message, that my kernel wasn't compiled with the zram option (?????). Is it a matter of the antiX-base.iso or of the virtualbox method ?

I wanted zRam to start at bootup.
From my fallible recall:

• * I copied /usr/local/bin/zram to / etc/init.d/zram
  * Ensured owner and group were root:root
  * Ensured permissions were rwxr-xr-x
  * Ran the command update-rc.d zram defaults
  * Rebooted the system

A quick check can be made that it has loaded by looking at the size of the swap reported by conky (if you use conky of course). Alternatively, a command such as blkid -o list should list a /dev/zram0 mounted as swap

I've had a quick look at the file size and date of /usr/local/bin/zram in both antiX-base and antiX-full. They appear to be the same so although I tested only antiX-full, it may work in antiX-base.

I checked my installations both on the virtualbox and the snapshot.
Ownership is correct and permissions are set to exectuable to rwx|r..|r..
Guess this should be enough as only root should have the permission to start init scipts

Tried the version from the forum as well as the one from /usr/local/bin

The output of lsmod|grep zram is -- nothing more

I think this is a specific problem of the thin client (or the virtualbox method...)

acutally, I'm afraid this is a bit over my head and it's better not to spend more time on this issue. Anyway, the best way to solve this problem is probably to insert more RAM or at least a CF card __{{emoticon}}__

btw -- I removed conky completely to save some RAM....