> On Jul 11, 2014, at 2:18 PM, "Adam W. Dace" <[email protected]> wrote: > > I've actually been experimenting with this a bit myself. > > Over time I've been frustrated by the fact that Traffic Server doesn't seem > to want to use its RAM cache much, on a system with very little load. > > Here's my current cache.config: > url_regex=\.(js|gif|jpg|jpeg|png|svg|css|ico|bmp|swf|flv|html|pdf)$ > scheme=http ttl-in-cache=60d > url_regex=.* scheme=http revalidate=1d > > Does this really help? I'm not really the one to ask, though it does seem to > make things a bit faster...and my RAM usage is up.
I'm sure this helps, but a forward proxy you are clearly violating the protocol :). As for RAM cache, it's simply a mapping over disk blocks. I don't understand why it would not use all RAM you give it. > > I've also been documenting how I tune Traffic Server on a personal basis, > feel free to check it out at WebProxyCacheTuning. > At this point I'm pretty sure I'm done tweaking Traffic Server v5.0. > > Regards, > > Adam > > >> On Mon, Jul 7, 2014 at 2:25 PM, Leif Hedstrom <[email protected]> wrote: >> >> On Jul 7, 2014, at 12:59 PM, Luís Antonio Pereira <[email protected]> >> wrote: >> >> > Hi all >> > >> > I would like to improve the performance of my webcache, for example >> > for example implementing the pining/cache.config: >> > >> > On records.config: >> > CONFIG proxy.config.cache.permit.pinning INT 1 >> >> >> Before you go there, I’d seriously investigate your cache and content. Some >> thoughts >> >> 1. Is your cache not large enough? How large is your active data content? >> Disks are fairly cheap, pinning is not (looking at your rules, I think >> you’ll kill your performance completely). >> >> 2. Do you have adequate RAM ? >> >> 3. What evidence is there to make you think pinning actually makes things >> better? Maybe what you are really looking for is to allocate some amount of >> the disks for Facebook content only ? This can be done with volume.config / >> hosting.config. >> >> 4. Are you sure you are not mixing up the notion of Cache-Control: max-age >> with pinning? >> >> >> >> In almost all cases I’ve seen ATS use, the cache does a great job on its >> own. Messing with things like cache pinning ought to be a last resort IMO, >> for some very specific use cases. I don’t think this is one of those. >> >> Cheers, >> >> — Leif > > > > -- > ____________________________________________________________ > Adam W. Dace <[email protected]> > > Phone: (815) 355-5848 > Instant Messenger: AIM & Yahoo! IM - colonelforbin74 | ICQ - #39374451 > Microsoft Messenger - [email protected] > > Google Profile: https://plus.google.com/u/0/109309036874332290399/about
