On Saturday 20 Jul 2013 06:12:40 Dale wrote: > Bruce Hill wrote: > > On Fri, Jul 19, 2013 at 08:34:27PM -0500, Dale wrote: > >>> Stop using disk and build in RAM: > >>> > >>> tmpfs /var/tmp/portage tmpfs > >>> size=7000M,nr_inodes=1M 0 0 tmpfs /dev/shm > >>> tmpfs nodev,nosuid,noexec 0 0 > >>> > >>> workstation ~ # free -m > >>> > >>> total used free shared buffers > >>> cached > >>> > >>> Mem: 15798 3711 12087 0 0 > >>> 937 -/+ buffers/cache: 2772 13025 > >>> Swap: 8103 0 8103 > >> > >> He may not have enough to do that tho. Some folks only have 4Gbs or > >> less still. That won't be enough for LOo. Heck, my 16Gbs wasn't enough > >> at one time. I had to either let it be on HDD or set it to a higher > >> amount than the default half. > >> > >> I also tested the time difference once before, it didn't really make > >> much difference. It just saves wear on a drive is all. > >> > >> Dale > > > > If 16GB of RAM wasn't enough, ydiw. I've used that line of 7G forever, > > and run app-office/libreoffice, as well as firefox and some other big > > app (forget it's name) and _never_ had a problem. > > Well, a while back, OOo and LOo wanted more than 8Gbs. It wasn't my > need but what portage looked for. Then someone did some changes and > reduced that need and it worked. From my understanding, there was some > code clean up that helped in that. I think it looks for 6Gbs now. From > the ebuild: > > CHECKREQS_MEMORY="512M" > CHECKREQS_DISK_BUILD="6G" > > It used to be more than that. If it didn't have enough, it stopped. > Even when I would override that setting, it would still run out of space > more often than not. As a matter of fact, I still have the command in > my freq used commands file that I used to fix it: > > mount -t tmpfs -o size=12g tmpfs /var/tmp/portage
Does it stop dead or does it start to page into swap? -- Regards, Mick
signature.asc
Description: This is a digitally signed message part.
