Re: A way to update revalidated metadata

From: Henrik Nordstrom <hno@dont-contact.us>
Date: Sun, 11 Aug 2002 00:27:56 +0200

On Friday 09 August 2002 08.03, Terence Kelly wrote:

> I'm no expert, but I think that APPEND has a major
> impact on the semantics of write(). I think that
> APPEND gives a certain atomicity guarantee that
> might be important in the code you're reading.
> On Solaris, "man -s 2 write" says

Currently Squid does not care. It is there as the only way of writing
to a store object is by storeAppend(), and maybe once in a time Squid
reused the same filedescriptor for concurrent swapins of the same
object (large, multiple clients)..

Removing O_APPEND is safe.

Regards
Henrik
Received on Sat Aug 10 2002 - 17:37:28 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:16:03 MST