Re: svn commit: r305968 - head/etc/autofs

2016-09-20 Thread Slawa Olhovchenkov
On Mon, Sep 19, 2016 at 10:19:04PM -0600, Warner Losh wrote: > For MSDOS it's one thing and likely helps. But for ufs it doesn't > help. Soft updates already does an excellent job at reducing wear and > all async does is give added danger of dataloss. Linux distros don't > matter for UFS because d

Re: svn commit: r305968 - head/etc/autofs

2016-09-19 Thread Edward Tomasz NapieraƂa
Good point; fixed. On 0919T2219, Warner Losh wrote: > For MSDOS it's one thing and likely helps. But for ufs it doesn't > help. Soft updates already does an excellent job at reducing wear and > all async does is give added danger of dataloss. Linux distros don't > matter for UFS because different

Re: svn commit: r305968 - head/etc/autofs

2016-09-19 Thread Warner Losh
For MSDOS it's one thing and likely helps. But for ufs it doesn't help. Soft updates already does an excellent job at reducing wear and all async does is give added danger of dataloss. Linux distros don't matter for UFS because different design decisions have been made for ext[234] and UFS. Warner

Re: svn commit: r305968 - head/etc/autofs

2016-09-19 Thread Edward Tomasz Napierala
Mounting removable media async improves performance and reduces wear. And AFAIK that's what most Linux distros used to do. On 0919T2032, Ronald Klop wrote: > Hi, > > Your commit message says in words exactly what the diff says in code. > Would you like to elaborate on why the change is made? I'

Re: svn commit: r305968 - head/etc/autofs

2016-09-19 Thread Ronald Klop
Hi, Your commit message says in words exactly what the diff says in code. Would you like to elaborate on why the change is made? I'm curious because this could cause severe damage to the filesystem on unclean eject of the media. Regards, Ronald. On Mon, 19 Sep 2016 10:51:27 +0200, Edwa