Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-26 Thread Yann Ylavic
On Wed, Aug 26, 2015 at 3:32 AM, Daniel Ruggeri drugg...@primary.net wrote: On 8/25/2015 10:11 AM, Jim Jagielski wrote: Again, if the slotmem exists and is persisted, the assumption is that THAT is what the admin wants, and when Apache restarts, THAT is the running config they desire. If there

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Jim Jagielski
I see that, but the assumption is that the reason why the size would have changed is that someone edited the httpd.conf file and added another balancer. At that point, the more logical thing would be to wipe the slate clean (as we do in other places where certain values in the slotmem != what we

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Jim Jagielski
The whole intent of the dynamic capability of balancer manage, plus the persistence of it, is to allow changes to be done to the reverse proxy setup w/o requiring any changes to the config at all. It allows it to be reconfigured dynamically, either via the manager or some other mechanism. It's to

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Yann Ylavic
With this commit the overall size of the storage slot does not change, the slot is attached and if it has enough room (growth margin) for the new configuration, it is used (like with the balancer-manager), otherwise the restart fails. On Tue, Aug 25, 2015 at 3:15 PM, Yann Ylavic

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Jim Jagielski
On Aug 25, 2015, at 10:41 AM, Yann Ylavic ylavic@gmail.com wrote: I don't see the difference between the configuration via balancer-manager or file+graceful-restart, both can change the number of balancers/members AND the settings while children are running. I'd kind of agree if we

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Yann Ylavic
The advantage of using the balancer-manager to add a member is that failures won't kill httpd, whereas failing to add member or balancer via the configuration file (no space left) stops httpd (now documented). To avoid that, maybe could we use non-shared (per child) memory in this case, like in

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Jim Jagielski
When 1st configured, Apache sees how many balancers there are set in the config file. It then pads that amount by growth to allow for the to-be-added dynamic addition of new balancers via the balancer manager (ala adding members). The assumption is that if the actual config file is changed, then

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Yann Ylavic
I don't see the difference between the configuration via balancer-manager or file+graceful-restart, both can change the number of balancers/members AND the settings while children are running. I'd kind of agree if we were talking about non-graceful restarts or first startup, we could add a new

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Yann Ylavic
On Tue, Aug 25, 2015 at 6:16 PM, Jim Jagielski j...@jagunet.com wrote: I see that, but the assumption is that the reason why the size would have changed is that someone edited the httpd.conf file and added another balancer. That's the only way to add a balancer AFAICT, the manager allows to

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Jim Jagielski
On Aug 25, 2015, at 2:15 PM, Yann Ylavic ylavic@gmail.com wrote: On Tue, Aug 25, 2015 at 6:16 PM, Jim Jagielski j...@jagunet.com wrote: I see that, but the assumption is that the reason why the size would have changed is that someone edited the httpd.conf file and added another

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Jim Jagielski
I don't understand this at all... I add new balancers and members all the time and do restarts and power cycles and never lose anything. What do you mean by some sort of size check failing?? The size of the slot should *never* change. On Aug 21, 2015, at 8:34 AM, yla...@apache.org wrote:

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Jim Jagielski
How and when do ALIGNED_PROXY_BALANCER_SHARED_SIZE and conf-max_balancers change??? On Aug 21, 2015, at 8:34 AM, yla...@apache.org wrote: Author: ylavic Date: Fri Aug 21 12:34:02 2015 New Revision: 1696960 URL: http://svn.apache.org/r1696960 Log: mod_proxy_balancer: Fix balancers and

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Yann Ylavic
Well, if one adds a new balancer before restarting, at post_config time conf-balancers-nelts is +1 wrt the previous load... On Tue, Aug 25, 2015 at 2:46 PM, Jim Jagielski j...@jagunet.com wrote: But the number of balancers does not change yet... And when it is implemented the overall size of

RE: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Plüm , Rüdiger , Vodafone Group
To: dev@httpd.apache.org Subject: Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c How and when do ALIGNED_PROXY_BALANCER_SHARED_SIZE and conf-max_balancers change??? On Aug 21, 2015, at 8:34 AM, yla

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Yann Ylavic
At every startup, in balancer_post_config: conf-max_balancers = conf-balancers-nelts + conf-bgrowth; which means that if you add a balancer and restart, storage-create() (eg. slotmem_create) will try to find an existing slot with this ID and the exact ALIGNED_PROXY_BALANCER_SHARED_SIZE *

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Jim Jagielski
But the number of balancers does not change yet... And when it is implemented the overall size of the storage slot should *never* change. It must be constant, ala the scoreboard. On Aug 25, 2015, at 8:39 AM, Yann Ylavic ylavic@gmail.com wrote: At every startup, in balancer_post_config:

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Jim Jagielski
You mean via changing the config file? If so, then all bets are off. You change the config file and restart then Apache should start off w/ a clean slate. Even so, you need to handle freeing the old shared mem segment and creating a brand new one with the correct size... Trying to use the old one

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Yann Ylavic
Hm, what's the point of the balancers and balancer-workers growth margin then? The first time the configration is loaded we create enough room for adding new balancers/workers, and on (graceful) restart we wouldn't use this space? That's also how the balancer-manager works when adding a new worker

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-25 Thread Daniel Ruggeri
On 8/25/2015 10:11 AM, Jim Jagielski wrote: Again, if the slotmem exists and is persisted, the assumption is that THAT is what the admin wants, and when Apache restarts, THAT is the running config they desire. If there are changes to the reverse proxy setup, the assumption must be we are

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-24 Thread Yann Ylavic
On Fri, Aug 21, 2015 at 8:07 PM, Ruediger Pluem rpl...@apache.org wrote: On 08/21/2015 07:02 PM, Yann Ylavic wrote: On Fri, Aug 21, 2015 at 4:15 PM, Ruediger Pluem rpl...@apache.org wrote: Hm. Doesn't that mean that if I keep on adding stuff and do graceful restarts that at one point in

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-21 Thread Ruediger Pluem
On 08/21/2015 02:34 PM, yla...@apache.org wrote: Author: ylavic Date: Fri Aug 21 12:34:02 2015 New Revision: 1696960 URL: http://svn.apache.org/r1696960 Log: mod_proxy_balancer: Fix balancers and balancer members reuse on restart when new ones are added. PR 58024. Since

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-21 Thread Ruediger Pluem
On 08/21/2015 07:02 PM, Yann Ylavic wrote: On Fri, Aug 21, 2015 at 4:15 PM, Ruediger Pluem rpl...@apache.org wrote: Hm. Doesn't that mean that if I keep on adding stuff and do graceful restarts that at one point in time I will have used up all my growth and need to do a hard restart?

Re: svn commit: r1696960 - in /httpd/httpd/trunk: CHANGES docs/log-message-tags/next-number modules/proxy/mod_proxy_balancer.c

2015-08-21 Thread Yann Ylavic
On Fri, Aug 21, 2015 at 4:15 PM, Ruediger Pluem rpl...@apache.org wrote: Hm. Doesn't that mean that if I keep on adding stuff and do graceful restarts that at one point in time I will have used up all my growth and need to do a hard restart? Yes, the last restart would fail. If yes, it