Re: haproxy-1.8.0, sending a email-alert causes 100% cpu usage, FreeBSD 11.1

Le 28/11/2017 à 07:25, Willy Tarreau a écrit : Hi Pieter, On Mon, Nov 27, 2017 at 09:43:52PM +0100, PiBa-NL wrote: Hi List, I thought i 'reasonably' tested some of 1.8.0's options. Today i put it into 'production' on my secondary cluster node and notice it takes 100% cpu... G.. bad.

Re: cache issue : some data skipped in the middle of the content

On Tue, Nov 28, 2017 at 10:56:56AM +0100, Cyril Bonté wrote: > Hi all, > > Late this night, I had an issue I didn't met during my previous tests, > related to the cache feature : sometimes, I had some corrupted images in the > browser. > After adding some debug messages, it seems there's an

Re: [BUG] haproxy 1.8-last/master-worker/peers

On Tue, Nov 28, 2017 at 01:43:52PM +0100, Frederic Lecaille wrote: > On 11/28/2017 01:29 PM, Frederic Lecaille wrote: > There is a "/* fall through */" between PEER_SESS_ST_CONNECT and > PEER_SESS_ST_GETSTATUS cases in peer I/O handler code, so with a lock which > may be taken two times ;). >From

cache issue : some data skipped in the middle of the content

Hi all, Late this night, I had an issue I didn't met during my previous tests, related to the cache feature : sometimes, I had some corrupted images in the browser. After adding some debug messages, it seems there's an issue in cache_store_http_forward_data() when there is not enough contiguous

Re: [BUG] haproxy 1.8-last/master-worker/peers

On 11/28/2017 02:13 PM, Willy Tarreau wrote: On Tue, Nov 28, 2017 at 01:43:52PM +0100, Frederic Lecaille wrote: On 11/28/2017 01:29 PM, Frederic Lecaille wrote: There is a "/* fall through */" between PEER_SESS_ST_CONNECT and PEER_SESS_ST_GETSTATUS cases in peer I/O handler code, so with a lock

Re: haproxy-1.8.0, sending a email-alert causes 100% cpu usage, FreeBSD 11.1

Hi Christopher, On Tue, Nov 28, 2017 at 10:28:20AM +0100, Christopher Faulet wrote: > Here is a patch that should fix the deadlock. Could you confirm it fixes > your bug ? At least it works for me, previously a config with a mailer would immediately go to 100% CPU once reporting a down backend,

Re: [1.8] tune.ssl.cachesize

Hi William, Thank you very much, I've applied your patch and it fixed issue: ps aux | grep hap USER PID %CPU %MEMVSZ RSS TTY STAT START TIME COMMAND haproxy 37601 0.1 0.5 2509548 165352 ? Ss 11:35 0:00 /usr/sbin/haproxy -f /etc/haproxy/haproxy.cfg -p

Re: cache issue : some data skipped in the middle of the content

On Tue, Nov 28, 2017 at 11:45:59AM +0100, William Lallemand wrote: > You're absolutely right, we spot this one before but it seems the patch didn't > make it to the master. > > This should be specific to the cache, we don't rewind and store the headers > in the other filters. > > Patch

Re: [1.8] tune.ssl.cachesize

On Tue, Nov 28, 2017 at 11:17:31AM +0100, William Lallemand wrote: > Thanks for the report, the bug was elsewhere, in fact the shctx was > initialized > for each ssl bind. > > Patch attached. Applied, thanks. Willy

Re: [BUG] haproxy 1.8-last/master-worker/peers

Hi Willy, > Le 28 nov. 2017 à 07:33, Willy Tarreau a écrit : > > Hi Manu, > > On Mon, Nov 27, 2017 at 06:21:50PM +0100, Emmanuel Hocdet wrote: >> Hi Willy, >> >>> Le 18 nov. 2017 à 12:28, Willy Tarreau a écrit : >>> >>> Hi Manu, >>> >>> On Fri, Nov 17, 2017 at

RE: Does HAProxy 1.8.0 need new param vs 1.7.9

Hello Willy, Thank you for giving time after a large 1.8.0 release to provide us vectors on testing HAProxy work mode. This community's intensity and laser focus are a standout. Below is all the information you asked us to look at. By any chance, was there any critical change between

Re: Does HAProxy 1.8.0 need new param vs 1.7.9

Hi Hemant, On Tue, Nov 28, 2017 at 12:15:32PM -0600, Coscend@HAProxy wrote: > Thank you for giving time after a large 1.8.0 release to provide us vectors > on testing HAProxy work mode. This community's intensity and laser focus > are a standout. Below is all the information you asked us to

Re:Re: Re: haproxy with many high Send-Q state

Hi Lukas. I am confused about haproxy logging. Haproxy configuration manual says "An optional level can be specified to filter outgoing messages. By default, all messages are sent". I leave option level as default (log 127.0.0.1 local0), so all message should be sent to syslog. But I can

Re: [BUG] haproxy 1.8-last/master-worker/peers

On Tue, Nov 28, 2017 at 02:17:19PM +0100, Frederic Lecaille wrote: > On 11/28/2017 02:13 PM, Willy Tarreau wrote: > > On Tue, Nov 28, 2017 at 01:43:52PM +0100, Frederic Lecaille wrote: > > > On 11/28/2017 01:29 PM, Frederic Lecaille wrote: > > > There is a "/* fall through */" between

Re: [BUG] haproxy 1.8-last/master-worker/peers

On Tue, Nov 28, 2017 at 12:22:04PM +0100, Emmanuel Hocdet wrote: > ok, i should have something strange because it’s easy to reproduce in my > environnement. > > When i look lsof i see on master: > haproxy 21355 root4u IPv4 39007164 0t0 TCP 10.101.20.4:943 > (LISTEN) > it’s link

Re: [BUG] haproxy 1.8-last/master-worker/peers

On Tue, Nov 28, 2017 at 02:56:55PM +0100, William Lallemand wrote: > On Tue, Nov 28, 2017 at 12:22:04PM +0100, Emmanuel Hocdet wrote: > > ok, i should have something strange because it’s easy to reproduce in my > > environnement. > > > > When i look lsof i see on master: > > haproxy 21355 root

Re: Why HAProxy is not a web server?

On 27 November 2017 at 01:09, wrote: > Why HAProxy is not a web server? Because it's a load balancer. It talks to multiple other web servers, often called backends or origins, which provide the content for it to serve to consumers. HTH, J

Re: haproxy-1.8.0, sending a email-alert causes 100% cpu usage, FreeBSD 11.1

Hi Christopher / Willy, On Tue, Nov 28, 2017 at 10:28:20AM +0100, Christopher Faulet wrote: Here is a patch that should fix the deadlock. Could you confirm it fixes your bug ? Fix confirmed. Thanks, PiBa-NL / Pieter

RE: Does HAProxy 1.8.0 need new param vs 1.7.9

Dear Willy, Thank you for your prompt guidance. Result of your insights: SUCCESS in loading stats and other pages in both the use cases listed below: 1). Removed mailers section in 1.8.0. 2). Compiled 1.8-commit 8f42b19 ("BUG/MEDIUM: tcp-check: Don't lock the server in tcpcheck_main")

haproxy 1.8.0 and http/2

hi, i just want to thank the haproxy team for making such a great programme. i have updated our production server to haproxy 1.8.0 and now be able to support http/2 to clients via haproxy. no need to change the backend servers as they still run on http/1.1. thanxs to the haproxy team Markus

[PATCH] BUG/MINOR: when master-worker is in daemon mode, detach from tty

Hi List, Made a patch that makes the master-worker detach from tty when it is also combined with daemon mode to allow a script to start haproxy with daemon mode, closing stdout so the calling process knows when to stop reading from it and allow the master to properly daemonize. This is

[PATCH] BUG/MINOR: Check if master-worker pipe getenv succeeded, also allow pipe fd 0 as valid.

Hi List, Willy / Willliam, A patch i came up with that might make it a little 'safer' with regard to getenv and its return value or possible lack thereof.. I'm not sure it it will ever happen. But if it does it wont fail on a null pointer or empty string conversion to a long value.. Though a

Re: Does HAProxy 1.8.0 need new param vs 1.7.9

On Tue, Nov 28, 2017 at 01:16:28PM -0600, Coscend@HAProxy wrote: > Thank you for your prompt guidance. Result of your insights: SUCCESS in > loading stats and other pages in both the use cases listed below: > > 1). Removed mailers section in 1.8.0. > > 2). Compiled 1.8-commit 8f42b19

HLS loadbalancing

Good day HAproxy looks like a very nice product. Was wondering if we can use it for HLS server load balancing. HLS is basically http file download but chunked. Connection times are way longer that normal and persistence will definitely be a key factor. Load balancer will ensure that

Re: HAProxy 1.7.9 Not Capturing Application Session Cookie

Hi Hermant, On 27/11/2017 20:31, Coscend@HAProxy wrote: > > Hello Moemen, > >   > > Thank you and very thoughtful of you to educate us on how HAProxy > handles Websockets and logs cookies.  Guidance such as these have > helped us grow from a rank startup to offer SLA-based healthcare > services

Re: [1.8] tune.ssl.cachesize

Hi Willy, Thanks for your response. In change http://git.haproxy.org/?p=haproxy-1.8.git;a=commit;h=4f45bb9c461f462290b77bf2511badb7a4453c0a William made shctx more generic. Right now it's calculated: shctx = (struct shared_context *)mmap(NULL, sizeof(struct shared_context) + extra + (maxblocks *

Re: [RFC] Wireshark dissector for SPOP

On 11/20/2017 10:16 PM, my.card@web.de wrote: Hi Fred, Hi Danny, thanks for looking into this, I've fixed this issue (and introduced some new :-)) The attached dissector code parses all SPOP frames sent from contrib/spoa_sample and haproxy and handles some fragmentation scenarios. Kind

Re: [PATCH] BUG/MINOR: Check if master-worker pipe getenv succeeded, also allow pipe fd 0 as valid.

On Tue, Nov 28, 2017 at 11:58:50PM +0100, PiBa-NL wrote: > Hi List, Willy / Willliam, > Hi Pieter, > A patch i came up with that might make it a little 'safer' with regard > to getenv and its return value or possible lack thereof.. I'm not sure > it it will ever happen. But if it does it wont

Re: [PATCH] BUG/MINOR: when master-worker is in daemon mode, detach from tty

On Wed, Nov 29, 2017 at 12:05:43AM +0100, PiBa-NL wrote: > Hi List, > Hi Pieter, > Made a patch that makes the master-worker detach from tty when it is > also combined with daemon mode to allow a script to start haproxy with > daemon mode, closing stdout so the calling process knows when to