Re: stats socket over TCP - dangerous?

On Mon, Sep 21, 2015 at 10:34:17PM +, Thrawn wrote: > > Does this mean that the TCP port is dangerous, or that the multiple > > instances of the socket is dangerous?? > > What is the danger with it? > > I believe the danger is that it may be externally visible, allowing outsiders > to start an

1.6-dev6 postponed a little bit...

Hi, Thierry was working on making it possible to implement internal services in Lua using applets (similar to stats, peers or CLI), which I decided to delay to 1.7 because we didn't want to push the release date further despite the feature being very interesting (otherwise we do that with every in

Re: stats socket over TCP - dangerous?

> software that can't use Unix sockets, like Java Er...that's generally true (although there are libraries like junixsocket), and not to put you off, but why do you have Java talking to the HAProxy admin socket?

Re: stats socket over TCP - dangerous?

The loopback address is a good idea for allowing communication locally, with software that can't use Unix sockets, like Java. -- ✉ Eduard Martinescu | ✆ (585) 708-9685 | [image: http://www.salsalabs.com] - ignite action. fuel change. On Mon

Re: stats socket over TCP - dangerous?

> Does this mean that the TCP port is dangerous, or that the multiple > instances of the socket is dangerous?? > What is the danger with it? I believe the danger is that it may be externally visible, allowing outsiders to start and stop your servers, add new backends and redirect traffic to them,

Re: HAProxy children crash when peers enabled

Good to know that you were able to reproduce the same, I will wait for your response while I try to debug it myself as well. Thanks On Sep 22, 2015 1:26 AM, "Willy Tarreau" wrote: > Pradeep, I reproduce it now, thank you very much for your detailed > configuration and procedure. I have no idea wh

Re: HAProxy children crash when peers enabled

Pradeep, I reproduce it now, thank you very much for your detailed configuration and procedure. I have no idea why it doesn't work reliably with this test while it does work with a regular IP address, it could be an on-wire protocol issue, I'll wait for Emeric tomorrow in the morning and we'll chec

Re: HAProxy children crash when peers enabled

On Tue, Sep 22, 2015 at 12:48:02AM +0530, Pradeep Jindal wrote: > Sorry about that, I am sending you my complete configuration (anonimized) > via private email. Perfect, I received it, thakn you! I'm investigating now. Willy

stats socket over TCP - dangerous?

Hi, I am looking at pulling stuff out of the stats socket, but the documentation is unclear on this point: It is also possible to add multiple instances of the stats socket by repeating the line, and make them listen to a TCP port instead of a UNIX socket. This is never done by default because th

Re: HAProxy children crash when peers enabled

On Mon, Sep 21, 2015 at 11:55 PM, Willy Tarreau wrote: > On Mon, Sep 21, 2015 at 11:50:34PM +0530, Pradeep Jindal wrote: > > OK, I will check the points you mentioned. But isn't it interesting that > > the second node correctly registers the entry in it's stick table, it's > > just that it doesn'

Re: [ANNOUNCE] haproxy-1.6-dev5

On Mon, Sep 21, 2015 at 08:37:36PM +0200, PiBa-NL wrote: > Op 14-9-2015 om 23:46 schreef Willy Tarreau: > >On Mon, Sep 14, 2015 at 10:46:00PM +0200, Baptiste wrote: > >>The issue is related to the connect() function to establish the UDP > >>connection. > >>Currently, I use a sizeof() to get the le

Re: [ANNOUNCE] haproxy-1.6-dev5

Op 14-9-2015 om 23:46 schreef Willy Tarreau: On Mon, Sep 14, 2015 at 10:46:00PM +0200, Baptiste wrote: The issue is related to the connect() function to establish the UDP connection. Currently, I use a sizeof() to get the len of the address structure and Remi suggested to use get_addr_len() inst

Re: HAProxy children crash when peers enabled

On Mon, Sep 21, 2015 at 11:50:34PM +0530, Pradeep Jindal wrote: > OK, I will check the points you mentioned. But isn't it interesting that > the second node correctly registers the entry in it's stick table, it's > just that it doesn't get replicated to first node. That was my question! Your messa

Re: HAProxy children crash when peers enabled

OK, I will check the points you mentioned. But isn't it interesting that the second node correctly registers the entry in it's stick table, it's just that it doesn't get replicated to first node. PFA my haproxy.cfg using which I can still reproduce the same behaviour as I stated, I used "socat ope

Re: HAProxy children crash when peers enabled

On Mon, Sep 21, 2015 at 11:14:25PM +0530, Pradeep Jindal wrote: > Which version (or sha would be better) are you running? The latest git with the latest crash fix : HA-Proxy version 1.6-dev5-a68f76-30 2015/09/21 That's 30 patches after dev5, so pretty similar. > I am running > 1.6-dev5 tarball

Re: HAProxy children crash when peers enabled

Which version (or sha would be better) are you running? I am running 1.6-dev5 tarball (patched with SRV_ADMF_CMAINT patch & this peers segfault patch) downloaded from haproxy.org. Here's my haproxy -vv: HA-Proxy version 1.6-dev5-a02e8a6 2015/09/14 Copyright 2000-2015 Willy Tarreau Build options

Re: HAProxy children crash when peers enabled

On Mon, Sep 21, 2015 at 10:58:00PM +0530, Pradeep Jindal wrote: > As I mentioned in my earlier email, I have tried with single process > (nbproc 1) too. In multi-process mode, I bind (bind-process 1) the > frontend-backend pair (pair using the stick table) to process 1 and I also > have a special u

Re: HAProxy children crash when peers enabled

As I mentioned in my earlier email, I have tried with single process (nbproc 1) too. In multi-process mode, I bind (bind-process 1) the frontend-backend pair (pair using the stick table) to process 1 and I also have a special unix stats socket bound (process 1) to process 1 which is what I query on

Re: HAProxy children crash when peers enabled

On Mon, Sep 21, 2015 at 09:53:34PM +0530, Pradeep Jindal wrote: > Applied the patch to 1.6-dev5, using exactly same configuration as I sent > over earlier on both of my haproxy instances. stick table entries are being > synced from A->B but not B->A. I have tried using single process too. Any > ide

Re: HAProxy children crash when peers enabled

Applied the patch to 1.6-dev5, using exactly same configuration as I sent over earlier on both of my haproxy instances. stick table entries are being synced from A->B but not B->A. I have tried using single process too. Any idea what could be wrong here? - Pradeep Jindal On Mon, Sep 21, 2015 at 7

Re: Re: conn_cur counter seems wrong

On Mon, Sep 21, 2015 at 05:34:46PM +0200, Willy Tarreau wrote: > On Mon, Sep 21, 2015 at 04:26:51PM +0200, Remi Gacogne wrote: > > On 09/21/2015 02:57 PM, Willy Tarreau wrote: > > > > > > On Mon, Sep 21, 2015 at 10:28:09AM +, mlist wrote: > > >> We installed "HA-Proxy version 1.6-dev5-caa6a1b

10.1" HMI, Kisok

This is a multi-part message in MIME format. Thismessagecontainsgraphics.Ifyoudonotseethe=graphics,cl=ickheretoview.FormoredetailsinformationaboutZB10=1A,Pleasevisitourwebsite:www.ewinsonic.com88633704789=88633704722sa...@ewinsonic.com

Re: Re: conn_cur counter seems wrong

On Mon, Sep 21, 2015 at 04:26:51PM +0200, Remi Gacogne wrote: > On 09/21/2015 02:57 PM, Willy Tarreau wrote: > > > > On Mon, Sep 21, 2015 at 10:28:09AM +, mlist wrote: > >> We installed "HA-Proxy version 1.6-dev5-caa6a1b 2015/09/17", as you can > >> see conn_cur in stick table seems wrong. Al

Re: Re: conn_cur counter seems wrong

Hi Rémi, On Mon, Sep 21, 2015 at 04:26:51PM +0200, Remi Gacogne wrote: > On 09/21/2015 02:57 PM, Willy Tarreau wrote: > > > > On Mon, Sep 21, 2015 at 10:28:09AM +, mlist wrote: > >> We installed "HA-Proxy version 1.6-dev5-caa6a1b 2015/09/17", as you can > >> see conn_cur in stick table seems

Re: Re: conn_cur counter seems wrong

On 09/21/2015 02:57 PM, Willy Tarreau wrote: > > On Mon, Sep 21, 2015 at 10:28:09AM +, mlist wrote: >> We installed "HA-Proxy version 1.6-dev5-caa6a1b 2015/09/17", as you can see >> conn_cur in stick table seems wrong. Also with a single http access test, we >> get inizial correct conn_cur=1

RE: Re: conn_cur counter seems wrong

Yes, it is always about 2^32 - conn_cnt on each test we have done. Roberto -Original Message- From: Remi Gacogne [mailto:rgaco...@coredump.fr] Sent: lunedì 21 settembre 2015 16.27 To: Willy Tarreau ; mlist Cc: 'haproxy@formilux.org' Subject: Re: Re: conn_cur counter seems wrong On 09

Re: HAProxy children crash when peers enabled

Yes, I was driving so couldn't reply with deetz. Processes not using any peers section would die. On Sep 21, 2015 6:54 PM, "Willy Tarreau" wrote: > On Mon, Sep 21, 2015 at 03:20:14PM +0200, Emeric Brun wrote: > > On 09/21/2015 03:03 PM, Willy Tarreau wrote: > > > On Mon, Sep 21, 2015 at 06:25:39P

Re: HAProxy children crash when peers enabled

On Mon, Sep 21, 2015 at 03:20:14PM +0200, Emeric Brun wrote: > On 09/21/2015 03:03 PM, Willy Tarreau wrote: > > On Mon, Sep 21, 2015 at 06:25:39PM +0530, Pradeep Jindal wrote: > >> I will give you more info in a while. I did some debugging, > >> "peers->sync_task = NULL;" seems to be the culprit he

Re: HAProxy children crash when peers enabled

On 09/21/2015 03:03 PM, Willy Tarreau wrote: > On Mon, Sep 21, 2015 at 06:25:39PM +0530, Pradeep Jindal wrote: >> I will give you more info in a while. I did some debugging, >> "peers->sync_task = NULL;" seems to be the culprit here, First, >> peers->sync_task is being set to NULL and then being de

Re: HAProxy children crash when peers enabled

On Mon, Sep 21, 2015 at 06:25:39PM +0530, Pradeep Jindal wrote: > I will give you more info in a while. I did some debugging, > "peers->sync_task = NULL;" seems to be the culprit here, First, > peers->sync_task is being set to NULL and then being deleted & freed which > I think would obviously fail

Re: conn_cur counter seems wrong

On Mon, Sep 21, 2015 at 10:28:09AM +, mlist wrote: > We installed "HA-Proxy version 1.6-dev5-caa6a1b 2015/09/17", as you can see > conn_cur in stick table seems wrong. Also with a single http access test, we > get inizial correct conn_cur=1 but after expired haproxy insert a new record > wit

Re: HAProxy children crash when peers enabled

I will give you more info in a while. I did some debugging, "peers->sync_task = NULL;" seems to be the culprit here, First, peers->sync_task is being set to NULL and then being deleted & freed which I think would obviously fail. I moved "peers->sync_task = NULL;" to after task_free call in the code

Re: HAProxy children crash when peers enabled

On Mon, Sep 21, 2015 at 11:50:27AM +0530, Pradeep Jindal wrote: > Hi, > > Please find the attached core dumps & config file. I have tried dev4 too, > same result. Please could you give a bit more context : - does it crash during startup or while running ? - in the latter case, how long does i

conn_cur counter seems wrong

We installed "HA-Proxy version 1.6-dev5-caa6a1b 2015/09/17", as you can see conn_cur in stick table seems wrong. Also with a single http access test, we get inizial correct conn_cur=1 but after expired haproxy insert a new record with same expire time with conn_cur like that below. So no more sr

C'est encore la rentrée des rencontres amoureuses

Title: La rentrée des rencontres Site de rencontre généraliste où les femmes ont le pouvoir Découvrez le site de rencontre MecACroquer ! Un site qui change les codes de la rencontre. Ce sont les femmes qui ont le pou