Re:Re: A bug? Puzzled at the actions after cache hit. :-(

Not being what it should, something wrong/unexpected/invalid.

Regards
Henrik

On Saturday 11 May 2002 16:06, maer727@sohu.com wrote:
> Thanks, Henrik pal!
>
> What means "cache is being corrupted" in your reply?
>
> Best regards,
> George Ma
>
> ----- Original Message -----
> From: Henrik Nordstrom
> To: maer727@sohu.com
> Cc: squid-dev@squid-cache.org
> Subject: Re: A bug? Puzzled at the actions after cache hit. :-(
> Sent: Sat May 11 20:42:20 CST 2002
>
> > The only case where the swap file size will change in between
> > where Squid wrote the file out and it is being read back again is
> > when the cache is being corrupted.
> >
> > Nothing but Squid is supposed to ever touch the cache files.
> > These are internal to Squid.
> >
> > Regards
> > Henrik
> >
> > On Saturday 11 May 2002 13:37, maer727@sohu.com wrote:
> > > Hi, pals!
> > >
> > > I think when cache hit, the function storeSwapMetaUnpack() will
> > > finally be called to change meta data to TVL list. I take a
> > > look at the function, storeClientReadHeader will finally be
> > > called. In this function, there is a statement like this,
> > >
> > > mem->object_sz = e->swap_file_sz - swap_hdr_sz;
> > >
> > > I think there may be a bug. Since the swap_file_sz of
> > > StoreEntry will possibly be changed since last time. IMHO, I
> > > think squid will not refresh the file size of the swap file
> > > since last time. I think we should force squid to re-read the
> > > file size of the swap file. Am I correct? ( Maybe some other
> > > process has modified the file and changed the size of the file.
> > > )
> > >
> > > Best regards,
> > > George Ma
Received on Sat May 11 2002 - 15:50:19 MDT