>> >> 2. added elevator=noop as a boot parameter >> > >> > I remember that I've given this second advice. Since then I've read in >> > the German computer journal c't [1] that CFQ has a detection for SSDs >> > since 2.6.28 and now is the best choice for these devices. >> >> OK, do I need a boot parameter if I've set CFQ as the default IO >> scheduler in the kernel config? > > No, that's what default switch is there for. > > >> > Yup, the entry should read: >> > tmp /tmp tmpfs default 0 0 > > I'd also suggest to explicitly specify max size of tmpfs mount, since > system locking because of wrong cp command is probably the last thing > you want. Argument is "size=" (see man 8 mount). > > >> Do you think mounting /tmp in RAM is worthwhile? Mike doesn't seem to >> think too highly of it. > > I guess accelerated fsync and reduced disk wear should be a nice plus > for SSD device, provided the path in question is constanly used for > writing which really might be the case with files, created in /tmp by > various mktemp implementations (like python's) which I haven't really > thought about, so I think I might be wrong about the issue here, sorry.
Thanks guys. I think the /tmp trick made a good difference. The last thing I can think of is pruning the kernel way down. I think it's mostly default. - Grant
