<div dir="ltr">On Mon, Oct 14, 2013 at 11:42 AM, Michael J. Hammel <span dir="ltr"><<a href="mailto:mjhammel@graphics-muse.org" target="_blank">mjhammel@graphics-muse.org</a>></span> wrote:<br><div class="gmail_extra">
<div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im"><span style="color:rgb(34,34,34)">nolock is fine if you're at home and the only one writing to a file.</span></div>
</blockquote><div><br></div><div>It's also useful when the filesytems you are mounting are for things that have their own locking systems and the ones that NFS implements break things. SVN is the big example I've run into often. Or things that are then exported for use by Samba.</div>
<div><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
tcp can slow accesses a little but deals with recovery better.</blockquote><div><br></div><div>The difference in my testing was negligably small. I really couldn't tell the difference. I imagine it mattered more back in pre-switched networking days?</div>
<div><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> Don't<br>
remember what async does.</blockquote><div><br></div><div>It lets the client assume that the writes are done before the server explicitly tells it they are. Helps prevent stuck mounts, and also why I included the tcp flag.</div>
<div><br></div><div>QH</div></div></div></div>