Samba logs and locks in tmpfs filesystem (Solaris)?

[Tristan Ball]

>It crossed my mind that all this "tdb" (or "STATUS..LCK") activity, and,
>indeed, even the logging activity, doesn't actually have to be done on a
>real disk-based filesystem.  It could actually be done in a "tmpfs"
>filesystem (the Solaris name for a pseudo-filesystem based in memory/swap,
>typically used for "/tmp" itself).
>
>Can anyone see any drawbacks?  I have had it running for a few hours now
>on our main fileserver, and it seems to have eased things.  (The only
>thing that I can see that we would lose is survival of log files across a
>reboot, but by a suitable "log file" I could adjust that...)


I think having the tdb files on a tmpfs filesystem is a great idea, and I 
think I'll trial it here. However I'm really not sure about the wisdom of 
putting logs on tmpfs as well, my reasons are as follows:

1) While the filesystem with your logs on it may be under a fairly high 
load, I doubt it's going to be under actual contention at the FS layer (I'm 
assuming seperate log files for each client?).
2) Depending on your debug level, it would be very easy to accidentally 
fill /tmp
3) I _like_ keeping logs across a reboot! :-)

I am very interested in your results tho, as you appear to be running a 
system that is very close to a big brother version of ours: We have E420r, 
2CPU's,2G ram and currently 639 smb processes, although these are split 
between a 4 installations of samba, 3 running on virtual interfaces, which 
provide server names and paths consistant with legacy NT servers, as we are 
in the middle of migrating/restructuring the network. The most connections 
any one virtual server gets is about 400.

T.