Re: Runaway process

2019-07-11 Thread Sander Klein
On 2019-07-12 04:27, Willy Tarreau wrote: If you can at least show the backtrace, this could be useful and we can see if the core would be needed or not. Maybe this will match another known bug. This is the BT of yesterday: --- GNU gdb (Debian 7.12-6) 7.12.0.20161007-git Copyright (C) 2016 Fr

Re: Runaway process

2019-07-11 Thread Willy Tarreau
Hi Sander, On Thu, Jul 11, 2019 at 01:28:50PM +0200, Sander Klein wrote: > On 2019-07-11 12:27, Tim Düsterhus wrote: > > Try attaching to the process with `gdb -p 12345` with 12345 being the > > process ID. Then: > > > > 1. Get a backtrace for all threads: thread apply all bt > > 2. Generate a co

Re: Runaway process

2019-07-11 Thread Sander Klein
On 2019-07-11 12:27, Tim Düsterhus wrote: Try attaching to the process with `gdb -p 12345` with 12345 being the process ID. Then: 1. Get a backtrace for all threads: thread apply all bt 2. Generate a core file: generate-core-file If you are also able to connect to the stats socket of that proce

Re: Runaway process

2019-07-11 Thread Tim Düsterhus
Sander, Am 11.07.19 um 08:48 schrieb Sander Klein: > I seem to have runaway HAProxy process since yesterday evening around > 20:50. This process is eating up 100% CPU continously. (HAProxy 1.9.8) > > Of course I can just kill it and go on with my life, but I was wondering > if there was any inter

Runaway process

2019-07-10 Thread Sander Klein
Hi, I seem to have runaway HAProxy process since yesterday evening around 20:50. This process is eating up 100% CPU continously. (HAProxy 1.9.8) Of course I can just kill it and go on with my life, but I was wondering if there was any interest to see if we can uncover a bug here. If so, plea