lttng-dev Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Mathieu Desnoyers via lttng-dev <lttng-dev@lists.lttng.org>
To: Jonathan Rajotte <jonathan.rajotte-julien@efficios.com>
Cc: lttng-dev <lttng-dev@lists.lttng.org>
Subject: Re: 回复:回复: 回复: some tracepoints not exist in metadata?
Date: Thu, 2 Jul 2020 10:03:10 -0400 (EDT)	[thread overview]
Message-ID: <445531696.20175.1593698590049.JavaMail.zimbra@efficios.com> (raw)
In-Reply-To: <20200702124541.GF4207@joraj-alpa>

----- On Jul 2, 2020, at 8:45 AM, Jonathan Rajotte jonathan.rajotte-julien@efficios.com wrote:

> On Thu, Jul 02, 2020 at 08:38:57AM +0800, zhenyu.ren via lttng-dev wrote:
>> Thanks a lot. I will try single quotes but it may take hours get the result.
>> Yesterday, I enabled the verbose option ,and found the following message "UST
>> app reply event failed. Application died (in add_event_ust_registry() at
>> ust-app.c:5405)" and "UST app notify socket unregister 34 (in
>> ust_app_notify_sock_unregister() at ust-app.c:5562)". In fact ,the socket that
>> Lttng-sessiond send reply to my app return an error with EPIPE. It seems that
>> the pipe breaked.So do you have any hint for that?
> 
> From lttng master:
>    ret = ustctl_reply_register_event(sock, event_id, ret_code);
>    if (ret < 0) {
>    	if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) {
>    		ERR("UST app reply event failed with ret %d", ret);
>    	} else {
>    		DBG3("UST app reply event failed. Application died");
>    	}
>    	/*
>    	 * No need to wipe the create event since the application socket will
>    	 * get close on error hence cleaning up everything by itself.
>    	 */
>    	goto error;
>    }
> 
> EPIPE ifor this socket is considered as an expected scenario.
> 
> This normally indicates a short-lived application or that your application
> crashed while lttng-ust and lttng-tools were performing an event registration.
> Considering that the path leading to the ustctl_reply_register_event call is
> preceded by a call fetching information from the same socket
> (ustctl_recv_register_event), I expect this to be a completely normal scenario
> unless you provide evidence that this application under tracing was acting
> normally and had a long lifetime.

In addition, one thing to consider is that if the application is a daemon
process closing all file descriptors, using fork, clone or daemon, then
helper libraries need to be preloaded with the daemon following instructions
in the man-page lttng-ust(3) under section "Using LTTng-UST with daemons".

Failure to preload those libraries in daemons would trigger the observed
scenario: the application hangs up the UST communication socket
and never reconnects.

Thanks,

Mathieu


-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com

WARNING: multiple messages have this Message-ID (diff)
From: Mathieu Desnoyers via lttng-dev <lttng-dev@lists.lttng.org>
To: Jonathan Rajotte <jonathan.rajotte-julien@efficios.com>
Cc: lttng-dev <lttng-dev@lists.lttng.org>
Subject: Re: [lttng-dev]  回复:回复: 回复: some tracepoints not exist in metadata?
Date: Thu, 2 Jul 2020 10:03:10 -0400 (EDT)	[thread overview]
Message-ID: <445531696.20175.1593698590049.JavaMail.zimbra@efficios.com> (raw)
Message-ID: <20200702140310.i44W8YQRT6khMAERpnwlYyaopsulEI8R1y9xTaJEuA0@z> (raw)
In-Reply-To: <20200702124541.GF4207@joraj-alpa>

----- On Jul 2, 2020, at 8:45 AM, Jonathan Rajotte jonathan.rajotte-julien@efficios.com wrote:

> On Thu, Jul 02, 2020 at 08:38:57AM +0800, zhenyu.ren via lttng-dev wrote:
>> Thanks a lot. I will try single quotes but it may take hours get the result.
>> Yesterday, I enabled the verbose option ,and found the following message "UST
>> app reply event failed. Application died (in add_event_ust_registry() at
>> ust-app.c:5405)" and "UST app notify socket unregister 34 (in
>> ust_app_notify_sock_unregister() at ust-app.c:5562)". In fact ,the socket that
>> Lttng-sessiond send reply to my app return an error with EPIPE. It seems that
>> the pipe breaked.So do you have any hint for that?
> 
> From lttng master:
>    ret = ustctl_reply_register_event(sock, event_id, ret_code);
>    if (ret < 0) {
>    	if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) {
>    		ERR("UST app reply event failed with ret %d", ret);
>    	} else {
>    		DBG3("UST app reply event failed. Application died");
>    	}
>    	/*
>    	 * No need to wipe the create event since the application socket will
>    	 * get close on error hence cleaning up everything by itself.
>    	 */
>    	goto error;
>    }
> 
> EPIPE ifor this socket is considered as an expected scenario.
> 
> This normally indicates a short-lived application or that your application
> crashed while lttng-ust and lttng-tools were performing an event registration.
> Considering that the path leading to the ustctl_reply_register_event call is
> preceded by a call fetching information from the same socket
> (ustctl_recv_register_event), I expect this to be a completely normal scenario
> unless you provide evidence that this application under tracing was acting
> normally and had a long lifetime.

In addition, one thing to consider is that if the application is a daemon
process closing all file descriptors, using fork, clone or daemon, then
helper libraries need to be preloaded with the daemon following instructions
in the man-page lttng-ust(3) under section "Using LTTng-UST with daemons".

Failure to preload those libraries in daemons would trigger the observed
scenario: the application hangs up the UST communication socket
and never reconnects.

Thanks,

Mathieu


-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com
_______________________________________________
lttng-dev mailing list
lttng-dev@lists.lttng.org
https://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev

  parent reply	other threads:[~2020-07-02 14:03 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-01  2:48 some tracepoints not exist in metadata? zhenyu.ren via lttng-dev
2020-07-01  2:48 ` [lttng-dev] " zhenyu.ren via lttng-dev
2020-07-01  3:43 ` 回复: " zhenyu.ren via lttng-dev
2020-07-01  3:43   ` [lttng-dev] " zhenyu.ren via lttng-dev
2020-07-01  3:56   ` 回复: " zhenyu.ren via lttng-dev
2020-07-01  3:56     ` [lttng-dev] " zhenyu.ren via lttng-dev
2020-07-01 12:07     ` Mathieu Desnoyers via lttng-dev
2020-07-01 12:07       ` [lttng-dev] " Mathieu Desnoyers via lttng-dev
2020-07-02  0:38       ` 回复:回复: " zhenyu.ren via lttng-dev
2020-07-02  0:38         ` [lttng-dev] " zhenyu.ren via lttng-dev
2020-07-02 12:45         ` Jonathan Rajotte-Julien via lttng-dev
2020-07-02 12:45           ` [lttng-dev] " Jonathan Rajotte-Julien via lttng-dev
2020-07-02 14:03           ` Mathieu Desnoyers via lttng-dev [this message]
2020-07-02 14:03             ` Mathieu Desnoyers via lttng-dev

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=445531696.20175.1593698590049.JavaMail.zimbra@efficios.com \
    --to=lttng-dev@lists.lttng.org \
    --cc=jonathan.rajotte-julien@efficios.com \
    --cc=mathieu.desnoyers@efficios.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox