X2Go Bug report logs - #1313
there is still a problem in getting a correct value for loadavgXX with loadchecker

Full log

🔗 View this message in rfc822 format

X-Loop: owner@bugs.x2go.org
Subject: Bug#1313: [X2Go-Dev] Bug#1313: there is still a problem in getting a correct value for loadavgXX with loadchecker
Reply-To: Mike Gabriel <mike.gabriel@das-netzwerkteam.de>, 1313@bugs.x2go.org
Resent-From: Mike Gabriel <mike.gabriel@das-netzwerkteam.de>
Resent-To: x2go-dev@lists.x2go.org
Resent-CC: X2Go Developers <x2go-dev@lists.x2go.org>
X-Loop: owner@bugs.x2go.org
Resent-Date: Thu, 13 Sep 2018 13:20:02 +0000
Resent-Message-ID: <handler.1313.B1313.15368447749152@bugs.x2go.org>
Resent-Sender: owner@bugs.x2go.org
X-X2Go-PR-Message: followup 1313
X-X2Go-PR-Package: x2gobroker-agent
X-X2Go-PR-Keywords: 
Received: via spool by 1313-submit@bugs.x2go.org id=B1313.15368447749152
          (code B ref 1313); Thu, 13 Sep 2018 13:20:02 +0000
Received: (at 1313) by bugs.x2go.org; 13 Sep 2018 13:19:34 +0000
X-Spam-Checker-Version: SpamAssassin 3.4.1 (2015-04-28) on
	ymir.das-netzwerkteam.de
X-Spam-Level: 
X-Spam-Status: No, score=-1.1 required=3.0 tests=BAYES_00,RDNS_NONE,
	URIBL_BLOCKED autolearn=no autolearn_force=no version=3.4.1
Received: from localhost (localhost [127.0.0.1])
	by ymir.das-netzwerkteam.de (Postfix) with ESMTP id 61B915DAEE
	for <1313@bugs.x2go.org>; Thu, 13 Sep 2018 15:19:31 +0200 (CEST)
X-Virus-Scanned: Debian amavisd-new at ymir.das-netzwerkteam.de
Received: from ymir.das-netzwerkteam.de ([127.0.0.1])
	by localhost (ymir.das-netzwerkteam.de [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id 0qh0wqBXy-6B for <1313@bugs.x2go.org>;
	Thu, 13 Sep 2018 15:19:25 +0200 (CEST)
Received: from fregna.das-netzwerkteam.de (unknown [IPv6:2a01:4f8:202:1381::1])
	by ymir.das-netzwerkteam.de (Postfix) with ESMTPS id 5AB025DA81
	for <1313@bugs.x2go.org>; Thu, 13 Sep 2018 15:19:25 +0200 (CEST)
Received: from grimnir.das-netzwerkteam.de (grimnir.das-netzwerkteam.de [IPv6:2a01:4f8:202:1381::105])
	by fregna.das-netzwerkteam.de (Postfix) with ESMTPS id 2A76D604F6;
	Thu, 13 Sep 2018 13:19:25 +0000 (UTC)
Received: from localhost (localhost [127.0.0.1])
	by grimnir.das-netzwerkteam.de (Postfix) with ESMTP id 26029C268C;
	Thu, 13 Sep 2018 15:19:25 +0200 (CEST)
X-Virus-Scanned: Debian amavisd-new at grimnir.das-netzwerkteam.de
Received: from grimnir.das-netzwerkteam.de ([127.0.0.1])
	by localhost (grimnir.das-netzwerkteam.de [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id 0wxeRlyXxcMo; Thu, 13 Sep 2018 15:19:19 +0200 (CEST)
Received: from das-netzwerkteam.de (localhost [127.0.0.1])
	by grimnir.das-netzwerkteam.de (Postfix) with ESMTPS id 6C76BC25C2;
	Thu, 13 Sep 2018 15:19:19 +0200 (CEST)
Received: from bifrost.das-netzwerkteam.de (bifrost.das-netzwerkteam.de
 [178.62.101.154]) by mail.das-netzwerkteam.de (Horde Framework) with HTTPS;
 Thu, 13 Sep 2018 13:19:19 +0000
Date: Thu, 13 Sep 2018 13:19:19 +0000
Message-ID: <20180913131919.Horde.E_bg6JMBLAcV9jdu5upQ4C5@mail.das-netzwerkteam.de>
From: Mike Gabriel <mike.gabriel@das-netzwerkteam.de>
To: Walid MOGHRABI <w.moghrabi@servicemagic.eu>, 1313@bugs.x2go.org
References: <883547417.4000531.1534156346103.JavaMail.root@servicemagic.eu>
 <1342096826.4000684.1534156440651.JavaMail.root@servicemagic.eu>
In-Reply-To: <1342096826.4000684.1534156440651.JavaMail.root@servicemagic.eu>
User-Agent: Horde Application Framework 5
Accept-Language: de,en
Organization: DAS-NETZWERKTEAM
X-Originating-IP: 178.62.101.154
X-Remote-Browser: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Firefox/52.0
Content-Type: multipart/signed; boundary="=_K-6vfqXkWCLYqM978CV7_VO";
 protocol="application/pgp-signature"; micalg=pgp-sha256
MIME-Version: 1.0

[Message part 1 (text/plain, inline)]

Hi Walid,

On  Mo 13 Aug 2018 12:34:00 CEST, Walid MOGHRABI wrote:

> package: x2gobroker-agent
> version: 0.0.4.0-0~1038~ubuntu16.04.1
> priority: bug
>
> I don't have a "0" value anymore since latest fixes so the  
> loadchecker process don't crash anymore but still, there is  
> something strange.
> Here is a fragment of my loadchecker logs from this morning.
> Just to give you the context, I have 22 servers which are all  
> automaticaly started at 6 AM (wake on lan) and they are absolutely  
> the same (blade servers with same CPU, memory amount, bios version,  
> ...).
> I checked our monitoring to see if users were correctly distributed  
> over the farm and at 7:30AM, I had about 7 or 8 users connected but  
> 4 of them were on tce-server-21 where I should have had 1 user on 8  
> servers.

Have you seen this issues more often? Does it hop from one server to  
another or occur on more than one server at a time?

> Here is the loadchecker log fragment :
>
> root@tce-manager-01 [~] # grep -B 1 'loadavgXX:1;'  
> /var/log/x2gobroker/loadchecker.log
> ...
> 2018-07-24 07:15:01,200 - loadchecker - INFO - Executing agent  
> command on remote host tce-server-21 (10.50.0.221): sh -c  
> '/usr/lib/x2go/x2gobroker-agent foo checkload'
> 2018-07-24 07:15:01,622 - loadchecker - INFO - Broker agent  
> answered: OK; loadavgXX:1; memAvail:23684; myMemAvail:23810;  
> numCPU:16; typeCPU:2400;
> --
> 2018-07-24 07:17:50,354 - loadchecker - INFO - Executing agent  
> command on remote host tce-server-21 (10.50.0.221): sh -c  
> '/usr/lib/x2go/x2gobroker-agent foo checkload'
> 2018-07-24 07:17:50,779 - loadchecker - INFO - Broker agent  
> answered: OK; loadavgXX:1; memAvail:23686; myMemAvail:23812;  
> numCPU:16; typeCPU:2400;
> --
> 2018-07-24 07:20:32,550 - loadchecker - INFO - Executing agent  
> command on remote host tce-server-21 (10.50.0.221): sh -c  
> '/usr/lib/x2go/x2gobroker-agent foo checkload'
> 2018-07-24 07:20:32,964 - loadchecker - INFO - Broker agent  
> answered: OK; loadavgXX:1; memAvail:23683; myMemAvail:23809;  
> numCPU:16; typeCPU:2400;
> --
> 2018-07-24 07:23:21,610 - loadchecker - INFO - Executing agent  
> command on remote host tce-server-21 (10.50.0.221): sh -c  
> '/usr/lib/x2go/x2gobroker-agent foo checkload'
> 2018-07-24 07:23:22,034 - loadchecker - INFO - Broker agent  
> answered: OK; loadavgXX:1; memAvail:23685; myMemAvail:23811;  
> numCPU:16; typeCPU:2400;
> --
> 2018-07-24 07:26:03,872 - loadchecker - INFO - Executing agent  
> command on remote host tce-server-21 (10.50.0.221): sh -c  
> '/usr/lib/x2go/x2gobroker-agent foo checkload'
> 2018-07-24 07:26:04,286 - loadchecker - INFO - Broker agent  
> answered: OK; loadavgXX:1; memAvail:23684; myMemAvail:23809;  
> numCPU:16; typeCPU:2400;
> --
> 2018-07-24 07:28:52,917 - loadchecker - INFO - Executing agent  
> command on remote host tce-server-21 (10.50.0.221): sh -c  
> '/usr/lib/x2go/x2gobroker-agent foo checkload'
> 2018-07-24 07:28:53,338 - loadchecker - INFO - Broker agent  
> answered: OK; loadavgXX:1; memAvail:23684; myMemAvail:23809;  
> numCPU:16; typeCPU:2400;
> --
> 2018-07-24 07:31:35,252 - loadchecker - INFO - Executing agent  
> command on remote host tce-server-21 (10.50.0.221): sh -c  
> '/usr/lib/x2go/x2gobroker-agent foo checkload'
> 2018-07-24 07:31:35,670 - loadchecker - INFO - Broker agent  
> answered: OK; loadavgXX:1; memAvail:23685; myMemAvail:23811;  
> numCPU:16; typeCPU:2400;
> --
> 2018-07-24 07:34:24,424 - loadchecker - INFO - Executing agent  
> command on remote host tce-server-21 (10.50.0.221): sh -c  
> '/usr/lib/x2go/x2gobroker-agent foo checkload'
> 2018-07-24 07:34:24,842 - loadchecker - INFO - Broker agent  
> answered: OK; loadavgXX:1; memAvail:23683; myMemAvail:23809;  
> numCPU:16; typeCPU:2400;

The log message "Broker agent answered:" comes directly from X2Go  
Broker Agent. It is basically its raw output.

This means, that the flaw must be in x2gobroker-agent.pl on the remote  
X2Go Server. Or that the loadchecker stops querying the broker agent  
and re-uses old data.

Looking at x2gobroker-agent.pl: If we focus on the loadavgXX for now,  
we come to the conclusion, that the load was really "0" or it was  
negative (both gives us a loadavgXX value of "1". The value should  
normally be greater (system load of 1.0 brings a loadavgXX of 100).

Looking at x2gobroker.agent.py: As the values always change slightly,  
we can't say that Python provides us the same return result string all  
the time. The query to the broker agent must have happened.

We need to do more debugging if this issue reoccurs:

  * run '/usr/lib/x2go/x2gobroker-agent foo checkload' on the  
affected X2Go Server
    and see if the reported values match with what the load checker sees.

  * check if it is reoccuring on the same X2Go Server

  * if /usr/lib/x2go/x2gobroker-agent returns a load of zero,
    look at /proc/loadavg

  * and /proc/sys/vm/min_free_kbytes,
    /proc/meminfo
    /proc/cpuinfo

... and report all back here...

> As you can see, there is only 1 server with a loadavgXX = 1 (which  
> means that in fact, we got a zero value from the broker agent).
> This is not normal, at 7:34, there were 4 users already connected to  
> this server and most of my other servers were empty.

> Restarting x2gobroker-loadchecker service fixed the issue.

Considering the above analysis that the issue must come from  
x2gobroker-agent.pl, a restart of the loadchecker can in theory not  
solve such an issue.

Can you see the x2gobroker-agent.pl process appear and disappear in  
the process list on the remote X2Go Server? Or does it stay open, even  
zombied?

> I think there is a problem in retrieving this informations ... even  
> memAvail seem strange on this server to me ... with 4 connected  
> users, it should have been lower than that.

Hmmm... Ok... Maybe the wrong server got tested? Two identical IPs on  
the subnet?

> I also think the number of connected users should be taken into  
> account when calculating the load factor (maybe this is already the  
> case, not sure about that).

Yes, we take the number of sessions into account. But that is not  
provided by the broker agent, but is available in the X2Go Server  
database and queried from there.

Mike
-- 

DAS-NETZWERKTEAM
mike gabriel, herweg 7, 24357 fleckeby
mobile: +49 (1520) 1976 148
landline: +49 (4354) 8390 139

GnuPG Fingerprint: 9BFB AEE8 6C0A A5FF BF22  0782 9AF4 6B30 2577 1B31
mail: mike.gabriel@das-netzwerkteam.de, http://das-netzwerkteam.de

[Message part 2 (application/pgp-signature, inline)]

Send a report that this bug log contains spam.