Re: haproxy invoked oom-killer

2013-02-02 Thread Baptiste
Do you mean 1.4.2 or 1.4.22 ??? I have in mind a few sysctls like tcp_mem, tcp_rmem and tcp_wmem. Some admins think it will improve things if they setup huge numbers... Resulting in memory over usage when HAProxy box has to manage many conn/s. Baptiste On Sun, Feb 3, 2013 at 5:17 AM, Dusty Doris

Re: haproxy invoked oom-killer

2013-02-02 Thread Dusty Doris
1.4.2 I agree. I don't understand how haproxy was utilizing all the ram. I'm not actually certain it was. Could you elaborate on the sysctls comment? On Saturday, February 2, 2013, Baptiste wrote: > Hi, > > My guess is that you have configured too huge buffers in HAProxy on in > your TCP sys

Re: haproxy invoked oom-killer

2013-02-02 Thread Dusty Doris
Lukas from what I see it looks like swap was not being utilized. It was available but not used. On Saturday, February 2, 2013, Lukas Tribus wrote: > > Notice that having the HAproxy box swapping is a huge performance killer > and you absolutely do not want to do that, so apart from tuning/config

Consistent hashing based on cookie - across multiple HAProxy boxes

2013-02-02 Thread Alex Davies
Hi All, What is the best way to configure haproxy to hash based on an application cookie (such as PHPSESSID), in a way that is consistent (meaning multiple haproxy servers will route to the same backend), ideally including the ability to weight backends (the configuration would clearly have to be

Re: haproxy invoked oom-killer

2013-02-02 Thread Baptiste
Hi, My guess is that you have configured too huge buffers in HAProxy on in your TCP sysctls. There is no reason for HAProxy to use all the available memory, unless you're running a 1.5-dev with a memory leak (dev15 had some kind of leak, at least I've already seen it being oomed). Could you pleas

Re: path_beg routing question

2013-02-02 Thread Baptiste
Hi, If your URL is /upload?param=value, then I would use the acl 'path' directly. It would be more accurate. path_beg would match anything starting by /upload. Well, it depends on your upcoming needs. cheers On Sat, Feb 2, 2013 at 5:35 PM, Saul Waizer wrote: > Hello list, > > I am working on

RE: haproxy invoked oom-killer

2013-02-02 Thread Lukas Tribus
Notice that having the HAproxy box swapping is a huge performance killer and you absolutely do not want to do that, so apart from tuning/configuring the OOM killer you should track the issue down and avoid memory depletion and swapping in first place.

path_beg routing question

2013-02-02 Thread Saul Waizer
Hello list, I am working on routing an upload module that my application uses to a new backend, it seems simple enough but I would like to get some feedback. The application uri looks like this: My.app.com/upload?id=123&morestuff=haproxy+is+awesome This is the config I have in mind: acl uploa

Re: Can HAProxy use a response header value to redirect to a different URL?

2013-02-02 Thread Jonathan Matthews
On 1 February 2013 23:32, Robert Snyder wrote: [snip] > I know it cannot be the most efficient thing, but it would allow us to have > custom redirection that can be maintained by the owners of the backend > pages. I've implemented something vaguely similar in nginx, and it was a pain. Far too muc

Re: haproxy invoked oom-killer

2013-02-02 Thread Jonathan Matthews
On 1 February 2013 21:57, Dusty Doris wrote: > oom-killer just killed my haproxy instance. Anyone know if there is a way > to prioritize haproxy and have it get killed after something else? Or, any > tuning that might help. https://www.google.co.uk/search?q=linux+exclude+proces+from+oom-killer

Re: Comparison to nginx

2013-02-02 Thread Willy Tarreau
On Fri, Feb 01, 2013 at 12:49:59PM +, William Lewis wrote: > I couldn't agree more, but I'm really in need of more concrete reasons for > pushing back against this. It is very possible that nginx could do the job right for you. That said, the feedback we get from people who love haproxy for lo