Looks good to me too
-Rob
signature.asc
Description: This is a digitally signed message part
Henrik Nordstrom wrote:
tor 2009-07-16 klockan 23:15 +1200 skrev Amos Jeffries:
Right, following up on that I had a hunch and discovered that n_active
is already performing this duty. The bug is therefore in my earlier fix
using n_running as a base instead of n_active.
How does this new atta
On Thu, 2009-07-16 at 09:11 +0200, Henrik Nordstrom wrote:
> tor 2009-07-16 klockan 16:54 +1000 skrev Robert Collins:
>
> > I thought we put a unique id on the stateful helper protocol _way_ back.
> > Sigh :(
>
> No, I couldn't when I implemented the concurent request protocol so it
> only got do
tor 2009-07-16 klockan 23:15 +1200 skrev Amos Jeffries:
> Right, following up on that I had a hunch and discovered that n_active
> is already performing this duty. The bug is therefore in my earlier fix
> using n_running as a base instead of n_active.
>
> How does this new attached patch look t
Amos Jeffries wrote:
Robert Collins wrote:
On Wed, 2009-07-15 at 00:52 +0200, Henrik Nordstrom wrote:
Or 4, go back to don't strictly enforce the number of helpers?
+1
I don't know what this strictly-enforce thing is, but it sounds unneeded
as we used to fire up the right number of helpers a
tor 2009-07-16 klockan 16:54 +1000 skrev Robert Collins:
> I thought we put a unique id on the stateful helper protocol _way_ back.
> Sigh :(
No, I couldn't when I implemented the concurent request protocol so it
only got done for stateless helpers. iirc I got stuck in trying to
unwind some of th
On Thu, 2009-07-16 at 08:42 +0200, Henrik Nordstrom wrote:
> tor 2009-07-16 klockan 14:47 +1200 skrev Amos Jeffries:
>
> > And the sad reality in squid-3 with _one_ helper:
> >
> > squid-*--1-helper---* winbindd [state1, state2]
> > \---crash: all 2 of 1 helpers all RS pending s
tor 2009-07-16 klockan 14:47 +1200 skrev Amos Jeffries:
> And the sad reality in squid-3 with _one_ helper:
>
> squid-*--1-helper---* winbindd [state1, state2]
> \---crash: all 2 of 1 helpers all RS pending state.
You can't run NTLM with just one helper, not until both the foll
tor 2009-07-16 klockan 11:13 +1000 skrev Robert Collins:
> Note that with NTLM this is realistic - there's no need for multiple
> helpers as a single helper can serve many concurrent requests.
No longer the case as we no longer do challenge reuses and the
concurrent protocol is not implemented fo
On Thu, 16 Jul 2009 12:19:55 +1000, Robert Collins
wrote:
> On Thu, 2009-07-16 at 14:08 +1200, Amos Jeffries wrote:
>>
>> Both reconfigure and helper recovery use startHelpers() where the
>> limit
>> needs to take place.
>> The DOS bug fix broke *rotate* (reconfigure has an async step added by
>>
NOte that winbind has a hard coded limit that is by default very low.
Opening 2n ntlm_auth helpers may make things blow up in horrible ways.
Adrian
2009/7/16 Robert Collins :
> On Thu, 2009-07-16 at 14:08 +1200, Amos Jeffries wrote:
>>
>> Both reconfigure and helper recovery use startHelpers()
On Thu, 2009-07-16 at 14:08 +1200, Amos Jeffries wrote:
>
> Both reconfigure and helper recovery use startHelpers() where the
> limit
> needs to take place.
> The DOS bug fix broke *rotate* (reconfigure has an async step added by
> Alex
> that prevents it being a problem).
s/rotate/reconfigure th
On Thu, 16 Jul 2009 11:13:14 +1000, Robert Collins
wrote:
> On Thu, 2009-07-16 at 12:44 +1200, Amos Jeffries wrote:
>>
>>
>> Well, as I'm seeing it now. The rotate/reconfigure case is are where
>> you
>> both argue its reasonable to have more than N helpers running.
>
> Yes. Trivial example of
On Thu, 2009-07-16 at 12:44 +1200, Amos Jeffries wrote:
>
>
> Well, as I'm seeing it now. The rotate/reconfigure case is are where
> you
> both argue its reasonable to have more than N helpers running.
Yes. Trivial example of problems: assume that the user configures one
stateful helper. They th
On Thu, 16 Jul 2009 09:15:22 +1000, Robert Collins
wrote:
> On Thu, 2009-07-16 at 01:19 +1200, Amos Jeffries wrote:
>
>
>> A Henrik said,
>>people with large memory-hog helpers have issues when Squid allocates
>> more than N bunches of their carefully tuned available memory to its
>> helpe
On Thu, 2009-07-16 at 01:19 +1200, Amos Jeffries wrote:
> A Henrik said,
>people with large memory-hog helpers have issues when Squid allocates
> more than N bunches of their carefully tuned available memory to its
> helpers. This is also important in low-memory systems requiring auth.
>
>
Bundle Buggy has detected this merge request.
For details, see:
http://bundlebuggy.aaronbentley.com/project/squid/request/%3C4A584880.2080707%40treenet.co.nz%3E
Project: Squid
Robert Collins wrote:
On Wed, 2009-07-15 at 00:52 +0200, Henrik Nordstrom wrote:
Or 4, go back to don't strictly enforce the number of helpers?
+1
I don't know what this strictly-enforce thing is, but it sounds unneeded
as we used to fire up the right number of helpers anyway.
I stopped Squ
On Wed, 2009-07-15 at 00:52 +0200, Henrik Nordstrom wrote:
>
> Or 4, go back to don't strictly enforce the number of helpers?
+1
I don't know what this strictly-enforce thing is, but it sounds unneeded
as we used to fire up the right number of helpers anyway.
-Rob
signature.asc
Description: T
lör 2009-07-11 klockan 20:08 +1200 skrev Amos Jeffries:
> It's most visible in rotate because Squid is intended to keep running
> with a hot-swap of its logs. Previously the sequence was causing two
> full sets of helpers to be started, and a period of overlap before the
> async closure of the o
http://www.squid-cache.org/bugs/show_bug.cgi?id=2680
I've tracked this one down to the helper shutdown being async and used
from mainRotate() which is a sync operation.
We are probably hitting the same issue on reconfigure and shutdown.
It's most visible in rotate because Squid is intended to
21 matches
Mail list logo
