Office extension - Libreoffice for TC 5.X

[vinnie]

I use this script to activate swapfile on a hd partition:

Code: [Select]

#!/bin/sh

case $1 in
on)
tce-load -i ntfs-3g &>/dev/null
sudo /usr/local/bin/ntfs-3g /dev/sda5 /mnt/sda5 &&
sudo swapon /mnt/sda5/tc.swp
;;
off)
#!/bin/sh
sudo swapoff /mnt/sda5/tc.swp &&
umount /mnt/sda5
;;
*)
echo -e "uso: _swap [ on ] attiva lo swap file
[ off ] disattiva lo swap file"
esac
exit 0


[theYinYeti]

Thanks vinnie

Loading just the right extensions to still be able to compile, but *not* trigger X11 launch (bug with my wife's laptop's gfx chip) was tricky but I managed.
And indeed with 2GB RAM and real swap, although once I thought the laptop had hanged, compilation finally went further than it had before.
But the poor laptop was a bit overwhelmed by the whole process (console refresh at the rate of 3 text lines per second ), and it eventually switched itself off abruptly because of overheating! And would not restart :/
I think it will restart this evening, after having cooled for a full day, and I'm rather confident that I'll be able to finish compiling.

However, I wonder: The PC shut itself down abruptly; if I remember correctly, this laptop's FS is Ext3; can I just let "make" continue to do its job, or should I "make clean; make" again, in case a file got corrupted?

[bmarkus]

Usually you can continue make. But before doing anything, make fsck.ext3 on the partition to fix possible fs errors.

[theYinYeti]

Thanks bmarkus, I'll do that. If I can do something at all, that is…

[curaga]

Be happy you're not compiling Firefox, that thing requires about 8gb RAM, last I checked (with LTO and other optimizations on)

[bmarkus]

Chromium:

It takes about 8GB of swap file to link chromium and its tests. If you get an out-of-memory error during the final link, you will need to add swap space with swapon. It's recommended to have at least 4GB of memory available for building a statically linked debug build. Dynamic linking and/or building a release build lowers memory requirements. People with less than 8GB of memory may want to not build tests since they are quite large.

[theYinYeti]

Code: [Select]

Virtual memory exhausted: cannot allocate memory

I should be able to borrow a better laptop in a couple of days… I'll tell you then how things went.

[tinypoodle]

Code: [Select]

Virtual memory exhausted: cannot allocate memory

You might give it a try with much more swapspace available.

[theYinYeti]

Hello tinypoodle
I thought about that, but my disk space was already so low, that I was fearing a shortage of disk space for the creation of the object files :/
Well, I warned straight from the start that I was not well equiped for such a compilation But I'll keep trying while I have ideas
Do you know if it would be complicated to make one TinyCore computer the "master" of others computers on the LAN, thus borrowing their RAM and CPU?

[tinypoodle]

Hello tinypoodle
I thought about that, but my disk space was already so low, that I was fearing a shortage of disk space for the creation of the object files :/

In that case there is a chance to squeeze out some more memory capacity by increasing size of zram disk while avoiding increasing swapspace on persistent storage.

Do you know if it would be complicated to make one TinyCore computer the "master" of others computers on the LAN, thus borrowing their RAM and CPU?

You could do a forum search for "cluster", but not sure if that's worth the effort only to compile BloatyOff... errr LibreOffice.

[Lee]

In that case there is a chance to squeeze out some more memory capacity by increasing size of zram disk while avoiding increasing swapspace on persistent storage.

Hmmmm... how about using a compressed disk-based swap - like zswap, only bigger - and slower.   

(I was kidding.)

[tinypoodle]

Hmmmm... how about using a compressed disk-based swap - like zswap, only bigger - and slower.   

(I was kidding.)

I can recall having witnessed discussions about such more than once in the past; not sure if it was only in IRC channels or elsewhere as well.
My impression is that it would appear that the trade off of gaining storage space for CPU cycles is not worth it, as opposed to gaining RAM for CPU cycles (as is the case with e.g. zram). ("You can never have too much RAM")

[theYinYeti]

Morning,
I got the laptop with 4GB RAM for one evening. It wouldn't boot the USB stick
So I installed VirtualBox on Windows (hey, 3GB RAM+10GB swap is still better than 2GB RAM+2GB swap…) and installed TC on a new virtual machine, using CorePlus iso.
The USB stick was inserted, so that I could later copy my files over to the fake hard disk. TC_install correctly listed sda1 (swap), sda2 (previously formatted as ext3) and sdb1 (usb), and I selected sda2 for installation. TC_install even displayed that it was setting the boot flag on sda2, but it failed to install nonetheless. Never did I select sdb1, nor did I mount it.
Long story short: TC_install made something to my USB stick in the process for no reason, rendering it unreadable, and I'm worse than back to square one, rather back to square zero
Time was running short (I had to give back the laptop in the morning, with all traces of my activity removed), so this marks the end of my efforts; I'm out of options. I'm going back to TC 4.7.7.
Good luck to the next one

[vinnie]

darn!

[Juanito]

libreoffice posted

Interesting that on the same machine, compiling the same version of libreoffice with the same ./configure switches I got:

time make build 3h 14m 16s [32-bit core]
time make build 1h 16m 15s [corepure64]

..which I guess shows the advantages of being able to use all 8gb of ram on the machine in question