Re: [MERGE] 1/2: Populate cache_mem on disk hits

2009-08-25 Thread Amos Jeffries
Henrik Nordstrom wrote: lör 2009-08-15 klockan 13:26 +1200 skrev Amos Jeffries: Now closed it, but I wonder, does this close or simply reduce the impact of bug #7? Reduces it sligthly as a sideeffect, but I would not account for it.. Am I right in thinking the in-memory copy gets demoted at

Re: [MERGE] 1/2: Populate cache_mem on disk hits

2009-08-20 Thread Bundle Buggy
Bundle Buggy has detected this merge request. For details, see: http://bundlebuggy.aaronbentley.com/project/squid/request/%3C1250159729.26060.6.camel%40henriknordstrom.net%3E Project: Squid

Re: [MERGE] 1/2: Populate cache_mem on disk hits

2009-08-14 Thread Henrik Nordstrom
lör 2009-08-15 klockan 13:26 +1200 skrev Amos Jeffries: > Now closed it, but I wonder, does this close or simply reduce the impact > of bug #7? Reduces it sligthly as a sideeffect, but I would not account for it.. > Am I right in thinking the in-memory copy gets demoted at some later > point b

Re: [MERGE] 1/2: Populate cache_mem on disk hits

2009-08-14 Thread Amos Jeffries
Henrik Nordstrom wrote: tor 2009-08-13 klockan 13:40 -0600 skrev Alex Rousskov: Henrik, please commit if you do not hear any objections. I am looking forward to benchmarking this important change. Done. Regards Henrik Bah, recalled that there was a big about this too. Now closed it, but

Re: [MERGE] 1/2: Populate cache_mem on disk hits

2009-08-14 Thread Henrik Nordstrom
tor 2009-08-13 klockan 13:40 -0600 skrev Alex Rousskov: > Henrik, please commit if you do not hear any objections. I am looking > forward to benchmarking this important change. Done. Regards Henrik

Re: [MERGE] 1/2: Populate cache_mem on disk hits

2009-08-14 Thread Henrik Nordstrom
fre 2009-08-14 klockan 09:37 +0200 skrev Kinkie: > Even longer term I believe it would be nice to introduce a > memory-backed cache_dir, and the ability to define a hierarchy of > cache_dirs with the ability to move objects among them, so that > cache_mem would only become a "transit" area.. Yes

Re: [MERGE] 1/2: Populate cache_mem on disk hits

2009-08-14 Thread Kinkie
On Fri, Aug 14, 2009 at 6:10 AM, Amos Jeffries wrote: > Henrik Nordstrom wrote: >> >> fre 2009-08-14 klockan 11:23 +1200 skrev Amos Jeffries: >> >>> cache_mem is where the fastest HITS/sec come from. Trashing it will >>> impact the very upper req/sec of Squid. As will forcing everything to disk >>>

Re: [MERGE] 1/2: Populate cache_mem on disk hits

2009-08-13 Thread Amos Jeffries
Henrik Nordstrom wrote: fre 2009-08-14 klockan 11:23 +1200 skrev Amos Jeffries: cache_mem is where the fastest HITS/sec come from. Trashing it will impact the very upper req/sec of Squid. As will forcing everything to disk first. Then again with things going disk->memory (how soon Henrik?) it

Re: [MERGE] 1/2: Populate cache_mem on disk hits

2009-08-13 Thread Henrik Nordstrom
fre 2009-08-14 klockan 11:23 +1200 skrev Amos Jeffries: > cache_mem is where the fastest HITS/sec come from. Trashing it will > impact the very upper req/sec of Squid. As will forcing everything to > disk first. Then again with things going disk->memory (how soon Henrik?) > it may not be notice

Re: [MERGE] 1/2: Populate cache_mem on disk hits

2009-08-13 Thread Amos Jeffries
Kinkie wrote: On Thu, Aug 13, 2009 at 7:32 PM, Henrik Nordstrom wrote: tor 2009-08-13 klockan 13:35 +0200 skrev Kinkie: On Thu, Aug 13, 2009 at 12:38 PM, Henrik Nordstrom wrote: squid.conf option for tuning when to keep objects in memory. +1 for the change :) (adri pointed out on IRC that it

Re: [MERGE] 1/2: Populate cache_mem on disk hits

2009-08-13 Thread Kinkie
On Thu, Aug 13, 2009 at 7:32 PM, Henrik Nordstrom wrote: > tor 2009-08-13 klockan 13:35 +0200 skrev Kinkie: >> On Thu, Aug 13, 2009 at 12:38 PM, Henrik >> Nordstrom wrote: >> > squid.conf option for tuning when to keep objects in memory. >> >> +1 for the change :) >> (adri pointed out on IRC that i

Re: [MERGE] 1/2: Populate cache_mem on disk hits

2009-08-13 Thread Alex Rousskov
On Thu, 13 Aug 2009, Kinkie wrote: On Thu, Aug 13, 2009 at 12:38 PM, Henrik Nordstrom wrote: squid.conf option for tuning when to keep objects in memory. +1 for the change :) Henrik, please commit if you do not hear any objections. I am looking forward to benchmarking this important chan

Re: [MERGE] 1/2: Populate cache_mem on disk hits

2009-08-13 Thread Henrik Nordstrom
tor 2009-08-13 klockan 13:35 +0200 skrev Kinkie: > On Thu, Aug 13, 2009 at 12:38 PM, Henrik > Nordstrom wrote: > > squid.conf option for tuning when to keep objects in memory. > > +1 for the change :) > (adri pointed out on IRC that it may put more pressure on cache_mem, > so some tweaks may be wa

Re: [MERGE] 1/2: Populate cache_mem on disk hits

2009-08-13 Thread Kinkie
On Thu, Aug 13, 2009 at 12:38 PM, Henrik Nordstrom wrote: > squid.conf option for tuning when to keep objects in memory. +1 for the change :) (adri pointed out on IRC that it may put more pressure on cache_mem, so some tweaks may be warranted) Also, a possible improvement might be to use a fast r

[MERGE] 1/2: Populate cache_mem on disk hits

2009-08-13 Thread Henrik Nordstrom
squid.conf option for tuning when to keep objects in memory. Sponsored by: The Measurement Factory # Bazaar merge directive format 2 (Bazaar 0.90) # revision_id: hen...@henriknordstrom.net-20090813102833-\ # 32fj8b4jw78a2dph # target_branch: ../memcache # testament_sha1: e84887264655e8bf193f0012

[MERGE] 1/2: Populate cache_mem on disk hits

2009-08-13 Thread Henrik Nordstrom
Populate cache_mem again on disk cache hits, moving on-disk objects back into the hot object pool. Sponsored by: The Measurement Factory # Bazaar merge directive format 2 (Bazaar 0.90) # revision_id: hen...@henriknordstrom.net-20090813102809-\ # muiiiqhlgun9yi1w # target_branch: http://www.squid