On Thu, Feb 19, 2009 at 11:17:50AM +0100, Stefan (metze) Metzmacher wrote: > > commit c441f58dedc465f59060296815a0bc7f9aeb743f > > Author: todd stecher <[email protected]> > > Date: Tue Feb 17 16:16:35 2009 -0800 > > > > S3: Allow SMBD processes to survive in low memory condidtions > > > > This commit adds a configure argument which allows for setting > > MADV_PROTECT > > in the madvise() API. With this enabled the kernel won't kill SMBD when > > it's running low on memory. > > > > wouldn't it make sense to add this call to become_daemon() so that > nmbd and winbindd also get the protection. > > I'm also wondering if we really want that the child (per connection) > smbds are also protected.
Yep, I think at most the parent smbd should be protected if at all. We've had bad memleaks in the past and I'm sure in certain workloads we still have them, and the kernel should be given a chance to kill the main culprit. Having all smbd's protected this way is definitely wrong. Volker
pgpwnRO2upSgU.pgp
Description: PGP signature
