Memory Leak - 10gen MMS | Google Groups

The old Google Groups will be going away soon, but your browser is incompatible with the new version.
Message from discussion Memory Leak
Date: Mon, 23 Apr 2012 11:09:12 -0700 (PDT)
From: Andrew Merts <and...@aos.me>
To: 10gen-mms@googlegroups.com
Message-ID: <16217831.1105.1335204552281.JavaMail.geo-discussion-forums@vbez20>
In-Reply-To: <23077478.39.1335200537607.JavaMail.geo-discussion-forums@vbhe9>
References: <31867446.48.1330102398482.JavaMail.geo-discussion-forums@pbnv3>
 <23233426.1942.1335196042130.JavaMail.geo-discussion-forums@ynbi17>
 <4396435.403.1335196264023.JavaMail.geo-discussion-forums@vbcg6>
 <12445817.604.1335197615117.JavaMail.geo-discussion-forums@ynbr30>
 <30145932.718.1335198041377.JavaMail.geo-discussion-forums@vbx14>
 <2195867.25.1335200213406.JavaMail.geo-discussion-forums@vbfx1>
 <23077478.39.1335200537607.JavaMail.geo-discussion-forums@vbhe9>
Subject: Re: Memory Leak
MIME-Version: 1.0
Content-Type: multipart/mixed; 
	boundary="----=_Part_1103_31686791.1335204552278"

------=_Part_1103_31686791.1335204552278
Content-Type: multipart/alternative; 
	boundary="----=_Part_1104_17517867.1335204552279"

------=_Part_1104_17517867.1335204552279
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit

All I've got in terms of logs was through MMS which isn't going to be that 
helpful because the Agent was down. As far as other services running on the 
machine it's just running MySQL, Apache, Mongod, the agent process, and a 
couple of PHP scripts downloading images. The Agent was monitoring the 
server it was running on and also an additional server over a private 
interface between the two. Is there any way to share access to MMS? it has 
the records for when the agent went down and it looks like Munin was 
working it was just showing no data because the agent went down. I'm not 
seeing anything in the hardware graphs for memory but on the status page 
you can see big dips in the resident memory for Mongod.

On Monday, April 23, 2012 12:02:17 PM UTC-5, Ryan wrote:
>
>
> Ok, thanks for the information. If you have any generic system charts for 
> the machine, that would be helpful. If you are not interested in sharing 
> those on a public forum, you can create a private Jira ticket here:
>
>
> https://jira.mongodb.org/secure/CreateIssue.jspa?pid=10582&issuetype=7&Create=Create
>
> Do you have other processes running on the machine? How many hosts are you 
> monitoring?
>
> On Monday, April 23, 2012 12:56:53 PM UTC-4, Andrew Merts wrote:
>>
>> Unfortunately not, I couldn't get Munin to cooperate with MMS and it 
>> wasn't much of an issue. I can tell you that when I killed it Memory usage 
>> was 8GB in RAM and 16GB in swap and htop showed agentProcess.py using %90.
>>
>> It didn't look so much like it was leaking memory as it was just eating 
>> all of it every couple of weeks. Also, yesterday was the first time that I 
>> was able to see the issue, we've had this happening occasionally every 
>> couple of weeks where our server would just become completely unresponsive 
>> but after ~1.5 hours everything would go back to normal and we couldn't 
>> find anything wrong when we got back in. It looks just like the first post 
>> in this thread except the issue would only happen every couple of weeks and 
>> it would fix itself if left alone for long enough.
>>
>> On Monday, April 23, 2012 11:20:41 AM UTC-5, Ryan wrote:
>>>
>>>
>>> Sorry... this thread is scattered. Do you have any munin/hardware charts 
>>> that showcase that the agent is leaking memory? Thanks!
>>>
>>> On Monday, April 23, 2012 12:13:35 PM UTC-4, Andrew Merts wrote:
>>>>
>>>> I don't think I'm using PyPy if I'm reading the output of this right
>>>>
>>>>
>>>> Python 2.6.6 (r266:84292, Dec 26 2010, 22:31:48) 
>>>> [GCC 4.4.5] on linux2
>>>> Type "help", "copyright", "credits" or "license" for more information.
>>>> >>> import pymongo
>>>> >>> pymongo.has_c()
>>>> True
>>>> >>> pymongo.version
>>>> '2.1.1'
>>>>
>>>>
>>>> Also It appears to have the C extensions too, Has anyone else had this 
>>>> issue without using PyPy and with the C extensions?
>>>>
>>>>
>>>> On Monday, April 23, 2012 10:51:04 AM UTC-5, Ryan wrote:
>>>>>
>>>>>
>>>>> Hi Andrew,
>>>>>
>>>>> Technically, we do not support PyPy or modifications to the agent MMS. 
>>>>> We will update the docs to state this.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> -Ryan
>>>>>
>>>>> On Monday, April 23, 2012 11:47:22 AM UTC-4, Andrew Merts wrote:
>>>>>>
>>>>>> I can confirm that I am also seeing these issues and it would be nice 
>>>>>> if the documentation had a warning about it. I don't want my database 
>>>>>> server going down just because of a "lightweight" monitoring script.
>>>>>>
>>>>>> On Friday, February 24, 2012 10:53:18 AM UTC-6, John Nunemaker wrote:
>>>>>>>
>>>>>>> I installed mms on a slave database. From there it talks to both 
>>>>>>> master/slave. We have pretty consistently been seeing the mms-agent grab 
>>>>>>> almost all the available memory. Probably happens like 5-10 times a day.
>>>>>>>
>>>>>>> http://cl.ly/EV04
>>>>>>>
>>>>>>>
>>>>>>> Anyone else having issues with this? Anything else I can provide to 
>>>>>>> help track this down?
>>>>>>>
>>>>>>
>> On Monday, April 23, 2012 11:20:41 AM UTC-5, Ryan wrote:
>>>
>>>
>>> Sorry... this thread is scattered. Do you have any munin/hardware charts 
>>> that showcase that the agent is leaking memory? Thanks!
>>>
>>> On Monday, April 23, 2012 12:13:35 PM UTC-4, Andrew Merts wrote:
>>>>
>>>> I don't think I'm using PyPy if I'm reading the output of this right
>>>>
>>>>
>>>> Python 2.6.6 (r266:84292, Dec 26 2010, 22:31:48) 
>>>> [GCC 4.4.5] on linux2
>>>> Type "help", "copyright", "credits" or "license" for more information.
>>>> >>> import pymongo
>>>> >>> pymongo.has_c()
>>>> True
>>>> >>> pymongo.version
>>>> '2.1.1'
>>>>
>>>>
>>>> Also It appears to have the C extensions too, Has anyone else had this 
>>>> issue without using PyPy and with the C extensions?
>>>>
>>>>
>>>> On Monday, April 23, 2012 10:51:04 AM UTC-5, Ryan wrote:
>>>>>
>>>>>
>>>>> Hi Andrew,
>>>>>
>>>>> Technically, we do not support PyPy or modifications to the agent MMS. 
>>>>> We will update the docs to state this.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> -Ryan
>>>>>
>>>>> On Monday, April 23, 2012 11:47:22 AM UTC-4, Andrew Merts wrote:
>>>>>>
>>>>>> I can confirm that I am also seeing these issues and it would be nice 
>>>>>> if the documentation had a warning about it. I don't want my database 
>>>>>> server going down just because of a "lightweight" monitoring script.
>>>>>>
>>>>>> On Friday, February 24, 2012 10:53:18 AM UTC-6, John Nunemaker wrote:
>>>>>>>
>>>>>>> I installed mms on a slave database. From there it talks to both 
>>>>>>> master/slave. We have pretty consistently been seeing the mms-agent grab 
>>>>>>> almost all the available memory. Probably happens like 5-10 times a day.
>>>>>>>
>>>>>>> http://cl.ly/EV04
>>>>>>>
>>>>>>>
>>>>>>> Anyone else having issues with this? Anything else I can provide to 
>>>>>>> help track this down?
>>>>>>>
>>>>>>
On Monday, April 23, 2012 12:02:17 PM UTC-5, Ryan wrote:
>
>
> Ok, thanks for the information. If you have any generic system charts for 
> the machine, that would be helpful. If you are not interested in sharing 
> those on a public forum, you can create a private Jira ticket here:
>
>
> https://jira.mongodb.org/secure/CreateIssue.jspa?pid=10582&issuetype=7&Create=Create
>
> Do you have other processes running on the machine? How many hosts are you 
> monitoring?
>
> On Monday, April 23, 2012 12:56:53 PM UTC-4, Andrew Merts wrote:
>>
>> Unfortunately not, I couldn't get Munin to cooperate with MMS and it 
>> wasn't much of an issue. I can tell you that when I killed it Memory usage 
>> was 8GB in RAM and 16GB in swap and htop showed agentProcess.py using %90.
>>
>> It didn't look so much like it was leaking memory as it was just eating 
>> all of it every couple of weeks. Also, yesterday was the first time that I 
>> was able to see the issue, we've had this happening occasionally every 
>> couple of weeks where our server would just become completely unresponsive 
>> but after ~1.5 hours everything would go back to normal and we couldn't 
>> find anything wrong when we got back in. It looks just like the first post 
>> in this thread except the issue would only happen every couple of weeks and 
>> it would fix itself if left alone for long enough.
>>
>> On Monday, April 23, 2012 11:20:41 AM UTC-5, Ryan wrote:
>>>
>>>
>>> Sorry... this thread is scattered. Do you have any munin/hardware charts 
>>> that showcase that the agent is leaking memory? Thanks!
>>>
>>> On Monday, April 23, 2012 12:13:35 PM UTC-4, Andrew Merts wrote:
>>>>
>>>> I don't think I'm using PyPy if I'm reading the output of this right
>>>>
>>>>
>>>> Python 2.6.6 (r266:84292, Dec 26 2010, 22:31:48) 
>>>> [GCC 4.4.5] on linux2
>>>> Type "help", "copyright", "credits" or "license" for more information.
>>>> >>> import pymongo
>>>> >>> pymongo.has_c()
>>>> True
>>>> >>> pymongo.version
>>>> '2.1.1'
>>>>
>>>>
>>>> Also It appears to have the C extensions too, Has anyone else had this 
>>>> issue without using PyPy and with the C extensions?
>>>>
>>>>
>>>> On Monday, April 23, 2012 10:51:04 AM UTC-5, Ryan wrote:
>>>>>
>>>>>
>>>>> Hi Andrew,
>>>>>
>>>>> Technically, we do not support PyPy or modifications to the agent MMS. 
>>>>> We will update the docs to state this.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> -Ryan
>>>>>
>>>>> On Monday, April 23, 2012 11:47:22 AM UTC-4, Andrew Merts wrote:
>>>>>>
>>>>>> I can confirm that I am also seeing these issues and it would be nice 
>>>>>> if the documentation had a warning about it. I don't want my database 
>>>>>> server going down just because of a "lightweight" monitoring script.
>>>>>>
>>>>>> On Friday, February 24, 2012 10:53:18 AM UTC-6, John Nunemaker wrote:
>>>>>>>
>>>>>>> I installed mms on a slave database. From there it talks to both 
>>>>>>> master/slave. We have pretty consistently been seeing the mms-agent grab 
>>>>>>> almost all the available memory. Probably happens like 5-10 times a day.
>>>>>>>
>>>>>>> http://cl.ly/EV04
>>>>>>>
>>>>>>>
>>>>>>> Anyone else having issues with this? Anything else I can provide to 
>>>>>>> help track this down?
>>>>>>>
>>>>>>
>> On Monday, April 23, 2012 11:20:41 AM UTC-5, Ryan wrote:
>>>
>>>
>>> Sorry... this thread is scattered. Do you have any munin/hardware charts 
>>> that showcase that the agent is leaking memory? Thanks!
>>>
>>> On Monday, April 23, 2012 12:13:35 PM UTC-4, Andrew Merts wrote:
>>>>
>>>> I don't think I'm using PyPy if I'm reading the output of this right
>>>>
>>>>
>>>> Python 2.6.6 (r266:84292, Dec 26 2010, 22:31:48) 
>>>> [GCC 4.4.5] on linux2
>>>> Type "help", "copyright", "credits" or "license" for more information.
>>>> >>> import pymongo
>>>> >>> pymongo.has_c()
>>>> True
>>>> >>> pymongo.version
>>>> '2.1.1'
>>>>
>>>>
>>>> Also It appears to have the C extensions too, Has anyone else had this 
>>>> issue without using PyPy and with the C extensions?
>>>>
>>>>
>>>> On Monday, April 23, 2012 10:51:04 AM UTC-5, Ryan wrote:
>>>>>
>>>>>
>>>>> Hi Andrew,
>>>>>
>>>>> Technically, we do not support PyPy or modifications to the agent MMS. 
>>>>> We will update the docs to state this.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> -Ryan
>>>>>
>>>>> On Monday, April 23, 2012 11:47:22 AM UTC-4, Andrew Merts wrote:
>>>>>>
>>>>>> I can confirm that I am also seeing these issues and it would be nice 
>>>>>> if the documentation had a warning about it. I don't want my database 
>>>>>> server going down just because of a "lightweight" monitoring script.
>>>>>>
>>>>>> On Friday, February 24, 2012 10:53:18 AM UTC-6, John Nunemaker wrote:
>>>>>>>
>>>>>>> I installed mms on a slave database. From there it talks to both 
>>>>>>> master/slave. We have pretty consistently been seeing the mms-agent grab 
>>>>>>> almost all the available memory. Probably happens like 5-10 times a day.
>>>>>>>
>>>>>>> http://cl.ly/EV04
>>>>>>>
>>>>>>>
>>>>>>> Anyone else having issues with this? Anything else I can provide to 
>>>>>>> help track this down?
>>>>>>>
>>>>>>
------=_Part_1104_17517867.1335204552279
Content-Type: text/html; charset=utf-8
Content-Transfer-Encoding: quoted-printable

All I've got in terms of logs was through MMS which isn't going to be that =
helpful because the Agent was down. As far as other services running on the=
 machine it's just running MySQL, Apache, Mongod, the agent process, and a =
couple of PHP scripts downloading images. The Agent was monitoring the serv=
er it was running on and also an additional server over a private interface=
 between the two. Is there any way to share access to MMS? it has the recor=
ds for when the agent went down and it looks like Munin was working it was =
just showing no data because the agent went down. I'm not seeing anything i=
n the hardware graphs for memory but on the status page you can see big dip=
s in the resident memory for Mongod.<br><br>On Monday, April 23, 2012 12:02=
:17 PM UTC-5, Ryan wrote:<blockquote class=3D"gmail_quote" style=3D"margin:=
 0;margin-left: 0.8ex;border-left: 1px #ccc solid;padding-left: 1ex;"><div>=
<br></div><div>Ok, thanks for the information. If you have any generic syst=
em charts for the machine, that would be helpful. If you are not interested=
 in sharing those on a public forum, you can create a private Jira ticket h=
ere:</div><div><br></div><div><a href=3D"https://jira.mongodb.org/secure/Cr=
eateIssue.jspa?pid=3D10582&amp;issuetype=3D7&amp;Create=3DCreate" target=3D=
"_blank">https://jira.mongodb.org/<wbr>secure/CreateIssue.jspa?pid=3D<wbr>1=
0582&amp;issuetype=3D7&amp;Create=3D<wbr>Create</a><br></div><div><br></div=
><div>Do you have other processes running on the machine? How many hosts ar=
e you monitoring?</div><br>On Monday, April 23, 2012 12:56:53 PM UTC-4, And=
rew Merts wrote:<blockquote class=3D"gmail_quote" style=3D"margin:0;margin-=
left:0.8ex;border-left:1px #ccc solid;padding-left:1ex">Unfortunately not, =
I couldn't get Munin to cooperate with MMS and it wasn't much of an issue. =
I can tell you that when I killed it Memory usage was 8GB in RAM and 16GB i=
n swap and htop showed agentProcess.py using %90.<div><br></div><div>It did=
n't look so much like it was leaking memory as it was just eating all of it=
 every couple of weeks. Also, yesterday was the first time that I was able =
to see the issue, we've had this happening occasionally every couple of wee=
ks where our server would just become completely unresponsive but after ~1.=
5 hours everything would go back to normal and we couldn't find anything wr=
ong when we got back in. It looks just like the first post in this thread e=
xcept the issue would only happen every couple of weeks and it would fix it=
self if left alone for long enough.<br><br>On Monday, April 23, 2012 11:20:=
41 AM UTC-5, Ryan wrote:<blockquote class=3D"gmail_quote" style=3D"margin:0=
;margin-left:0.8ex;border-left:1px #ccc solid;padding-left:1ex"><div><br></=
div>Sorry... this thread is scattered. Do you have any munin/hardware chart=
s that showcase that the agent is leaking memory? Thanks!<br><br>On Monday,=
 April 23, 2012 12:13:35 PM UTC-4, Andrew Merts wrote:<blockquote class=3D"=
gmail_quote" style=3D"margin:0;margin-left:0.8ex;border-left:1px #ccc solid=
;padding-left:1ex">I don't think I'm using PyPy if I'm reading the output o=
f this right<div><br></div><div><br></div><div><div>Python 2.6.6 (r266:8429=
2, Dec 26 2010, 22:31:48)&nbsp;</div><div>[GCC 4.4.5] on linux2</div><div>T=
ype "help", "copyright", "credits" or "license" for more information.</div>=
<div>&gt;&gt;&gt; import pymongo</div><div>&gt;&gt;&gt; pymongo.has_c()</di=
v><div>True</div><div>&gt;&gt;&gt; pymongo.version</div><div>'2.1.1'</div><=
/div><div><br></div><div><br></div><div>Also It appears to have the C exten=
sions too, Has anyone else had this issue without using PyPy and with the C=
 extensions?</div><div><br><br>On Monday, April 23, 2012 10:51:04 AM UTC-5,=
 Ryan wrote:<blockquote class=3D"gmail_quote" style=3D"margin:0;margin-left=
:0.8ex;border-left:1px #ccc solid;padding-left:1ex"><div><br></div>Hi Andre=
w,<div><br></div><div>Technically, we do not support PyPy or modifications =
to the agent MMS. We will update the docs to state this.</div><div><br></di=
v><div>Thanks,</div><div><br></div><div>-Ryan<br><br>On Monday, April 23, 2=
012 11:47:22 AM UTC-4, Andrew Merts wrote:<blockquote class=3D"gmail_quote"=
 style=3D"margin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding-lef=
t:1ex">I can confirm that I am also seeing these issues and it would be nic=
e if the documentation had a warning about it. I don't want my database ser=
ver going down just because of a "lightweight" monitoring script.<br><br>On=
 Friday, February 24, 2012 10:53:18 AM UTC-6, John Nunemaker wrote:<blockqu=
ote class=3D"gmail_quote" style=3D"margin:0;margin-left:0.8ex;border-left:1=
px #ccc solid;padding-left:1ex">I installed mms on a slave database. From t=
here it talks to both master/slave. We have pretty consistently been seeing=
 the mms-agent grab almost all the available memory. Probably happens like =
5-10 times a day.<div><br></div><div><p style=3D"margin:0px 0px 0px 0px;fon=
t:12.0px Helvetica"><a href=3D"http://cl.ly/EV04" target=3D"_blank">http://=
cl.ly/EV04</a></p><p style=3D"margin:0px 0px 0px 0px;font:12.0px Helvetica"=
><br></p><p style=3D"margin:0px 0px 0px 0px;font:12.0px Helvetica">Anyone e=
lse having issues with this? Anything else I can provide to help track this=
 down?</p></div></blockquote></blockquote></div></blockquote></div></blockq=
uote></blockquote><br>On Monday, April 23, 2012 11:20:41 AM UTC-5, Ryan wro=
te:<blockquote class=3D"gmail_quote" style=3D"margin:0;margin-left:0.8ex;bo=
rder-left:1px #ccc solid;padding-left:1ex"><div><br></div>Sorry... this thr=
ead is scattered. Do you have any munin/hardware charts that showcase that =
the agent is leaking memory? Thanks!<br><br>On Monday, April 23, 2012 12:13=
:35 PM UTC-4, Andrew Merts wrote:<blockquote class=3D"gmail_quote" style=3D=
"margin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding-left:1ex">I =
don't think I'm using PyPy if I'm reading the output of this right<div><br>=
</div><div><br></div><div><div>Python 2.6.6 (r266:84292, Dec 26 2010, 22:31=
:48)&nbsp;</div><div>[GCC 4.4.5] on linux2</div><div>Type "help", "copyrigh=
t", "credits" or "license" for more information.</div><div>&gt;&gt;&gt; imp=
ort pymongo</div><div>&gt;&gt;&gt; pymongo.has_c()</div><div>True</div><div=
>&gt;&gt;&gt; pymongo.version</div><div>'2.1.1'</div></div><div><br></div><=
div><br></div><div>Also It appears to have the C extensions too, Has anyone=
 else had this issue without using PyPy and with the C extensions?</div><di=
v><br><br>On Monday, April 23, 2012 10:51:04 AM UTC-5, Ryan wrote:<blockquo=
te class=3D"gmail_quote" style=3D"margin:0;margin-left:0.8ex;border-left:1p=
x #ccc solid;padding-left:1ex"><div><br></div>Hi Andrew,<div><br></div><div=
>Technically, we do not support PyPy or modifications to the agent MMS. We =
will update the docs to state this.</div><div><br></div><div>Thanks,</div><=
div><br></div><div>-Ryan<br><br>On Monday, April 23, 2012 11:47:22 AM UTC-4=
, Andrew Merts wrote:<blockquote class=3D"gmail_quote" style=3D"margin:0;ma=
rgin-left:0.8ex;border-left:1px #ccc solid;padding-left:1ex">I can confirm =
that I am also seeing these issues and it would be nice if the documentatio=
n had a warning about it. I don't want my database server going down just b=
ecause of a "lightweight" monitoring script.<br><br>On Friday, February 24,=
 2012 10:53:18 AM UTC-6, John Nunemaker wrote:<blockquote class=3D"gmail_qu=
ote" style=3D"margin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding=
-left:1ex">I installed mms on a slave database. From there it talks to both=
 master/slave. We have pretty consistently been seeing the mms-agent grab a=
lmost all the available memory. Probably happens like 5-10 times a day.<div=
><br></div><div><p style=3D"margin:0px 0px 0px 0px;font:12.0px Helvetica"><=
a href=3D"http://cl.ly/EV04" target=3D"_blank">http://cl.ly/EV04</a></p><p =
style=3D"margin:0px 0px 0px 0px;font:12.0px Helvetica"><br></p><p style=3D"=
margin:0px 0px 0px 0px;font:12.0px Helvetica">Anyone else having issues wit=
h this? Anything else I can provide to help track this down?</p></div></blo=
ckquote></blockquote></div></blockquote></div></blockquote></blockquote></d=
iv></blockquote></blockquote><br>On Monday, April 23, 2012 12:02:17 PM UTC-=
5, Ryan wrote:<blockquote class=3D"gmail_quote" style=3D"margin: 0;margin-l=
eft: 0.8ex;border-left: 1px #ccc solid;padding-left: 1ex;"><div><br></div><=
div>Ok, thanks for the information. If you have any generic system charts f=
or the machine, that would be helpful. If you are not interested in sharing=
 those on a public forum, you can create a private Jira ticket here:</div><=
div><br></div><div><a href=3D"https://jira.mongodb.org/secure/CreateIssue.j=
spa?pid=3D10582&amp;issuetype=3D7&amp;Create=3DCreate" target=3D"_blank">ht=
tps://jira.mongodb.org/<wbr>secure/CreateIssue.jspa?pid=3D<wbr>10582&amp;is=
suetype=3D7&amp;Create=3D<wbr>Create</a><br></div><div><br></div><div>Do yo=
u have other processes running on the machine? How many hosts are you monit=
oring?</div><br>On Monday, April 23, 2012 12:56:53 PM UTC-4, Andrew Merts w=
rote:<blockquote class=3D"gmail_quote" style=3D"margin:0;margin-left:0.8ex;=
border-left:1px #ccc solid;padding-left:1ex">Unfortunately not, I couldn't =
get Munin to cooperate with MMS and it wasn't much of an issue. I can tell =
you that when I killed it Memory usage was 8GB in RAM and 16GB in swap and =
htop showed agentProcess.py using %90.<div><br></div><div>It didn't look so=
 much like it was leaking memory as it was just eating all of it every coup=
le of weeks. Also, yesterday was the first time that I was able to see the =
issue, we've had this happening occasionally every couple of weeks where ou=
r server would just become completely unresponsive but after ~1.5 hours eve=
rything would go back to normal and we couldn't find anything wrong when we=
 got back in. It looks just like the first post in this thread except the i=
ssue would only happen every couple of weeks and it would fix itself if lef=
t alone for long enough.<br><br>On Monday, April 23, 2012 11:20:41 AM UTC-5=
, Ryan wrote:<blockquote class=3D"gmail_quote" style=3D"margin:0;margin-lef=
t:0.8ex;border-left:1px #ccc solid;padding-left:1ex"><div><br></div>Sorry..=
. this thread is scattered. Do you have any munin/hardware charts that show=
case that the agent is leaking memory? Thanks!<br><br>On Monday, April 23, =
2012 12:13:35 PM UTC-4, Andrew Merts wrote:<blockquote class=3D"gmail_quote=
" style=3D"margin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding-le=
ft:1ex">I don't think I'm using PyPy if I'm reading the output of this righ=
t<div><br></div><div><br></div><div><div>Python 2.6.6 (r266:84292, Dec 26 2=
010, 22:31:48)&nbsp;</div><div>[GCC 4.4.5] on linux2</div><div>Type "help",=
 "copyright", "credits" or "license" for more information.</div><div>&gt;&g=
t;&gt; import pymongo</div><div>&gt;&gt;&gt; pymongo.has_c()</div><div>True=
</div><div>&gt;&gt;&gt; pymongo.version</div><div>'2.1.1'</div></div><div><=
br></div><div><br></div><div>Also It appears to have the C extensions too, =
Has anyone else had this issue without using PyPy and with the C extensions=
?</div><div><br><br>On Monday, April 23, 2012 10:51:04 AM UTC-5, Ryan wrote=
:<blockquote class=3D"gmail_quote" style=3D"margin:0;margin-left:0.8ex;bord=
er-left:1px #ccc solid;padding-left:1ex"><div><br></div>Hi Andrew,<div><br>=
</div><div>Technically, we do not support PyPy or modifications to the agen=
t MMS. We will update the docs to state this.</div><div><br></div><div>Than=
ks,</div><div><br></div><div>-Ryan<br><br>On Monday, April 23, 2012 11:47:2=
2 AM UTC-4, Andrew Merts wrote:<blockquote class=3D"gmail_quote" style=3D"m=
argin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding-left:1ex">I ca=
n confirm that I am also seeing these issues and it would be nice if the do=
cumentation had a warning about it. I don't want my database server going d=
own just because of a "lightweight" monitoring script.<br><br>On Friday, Fe=
bruary 24, 2012 10:53:18 AM UTC-6, John Nunemaker wrote:<blockquote class=
=3D"gmail_quote" style=3D"margin:0;margin-left:0.8ex;border-left:1px #ccc s=
olid;padding-left:1ex">I installed mms on a slave database. From there it t=
alks to both master/slave. We have pretty consistently been seeing the mms-=
agent grab almost all the available memory. Probably happens like 5-10 time=
s a day.<div><br></div><div><p style=3D"margin:0px 0px 0px 0px;font:12.0px =
Helvetica"><a href=3D"http://cl.ly/EV04" target=3D"_blank">http://cl.ly/EV0=
4</a></p><p style=3D"margin:0px 0px 0px 0px;font:12.0px Helvetica"><br></p>=
<p style=3D"margin:0px 0px 0px 0px;font:12.0px Helvetica">Anyone else havin=
g issues with this? Anything else I can provide to help track this down?</p=
></div></blockquote></blockquote></div></blockquote></div></blockquote></bl=
ockquote><br>On Monday, April 23, 2012 11:20:41 AM UTC-5, Ryan wrote:<block=
quote class=3D"gmail_quote" style=3D"margin:0;margin-left:0.8ex;border-left=
:1px #ccc solid;padding-left:1ex"><div><br></div>Sorry... this thread is sc=
attered. Do you have any munin/hardware charts that showcase that the agent=
 is leaking memory? Thanks!<br><br>On Monday, April 23, 2012 12:13:35 PM UT=
C-4, Andrew Merts wrote:<blockquote class=3D"gmail_quote" style=3D"margin:0=
;margin-left:0.8ex;border-left:1px #ccc solid;padding-left:1ex">I don't thi=
nk I'm using PyPy if I'm reading the output of this right<div><br></div><di=
v><br></div><div><div>Python 2.6.6 (r266:84292, Dec 26 2010, 22:31:48)&nbsp=
;</div><div>[GCC 4.4.5] on linux2</div><div>Type "help", "copyright", "cred=
its" or "license" for more information.</div><div>&gt;&gt;&gt; import pymon=
go</div><div>&gt;&gt;&gt; pymongo.has_c()</div><div>True</div><div>&gt;&gt;=
&gt; pymongo.version</div><div>'2.1.1'</div></div><div><br></div><div><br><=
/div><div>Also It appears to have the C extensions too, Has anyone else had=
 this issue without using PyPy and with the C extensions?</div><div><br><br=
>On Monday, April 23, 2012 10:51:04 AM UTC-5, Ryan wrote:<blockquote class=
=3D"gmail_quote" style=3D"margin:0;margin-left:0.8ex;border-left:1px #ccc s=
olid;padding-left:1ex"><div><br></div>Hi Andrew,<div><br></div><div>Technic=
ally, we do not support PyPy or modifications to the agent MMS. We will upd=
ate the docs to state this.</div><div><br></div><div>Thanks,</div><div><br>=
</div><div>-Ryan<br><br>On Monday, April 23, 2012 11:47:22 AM UTC-4, Andrew=
 Merts wrote:<blockquote class=3D"gmail_quote" style=3D"margin:0;margin-lef=
t:0.8ex;border-left:1px #ccc solid;padding-left:1ex">I can confirm that I a=
m also seeing these issues and it would be nice if the documentation had a =
warning about it. I don't want my database server going down just because o=
f a "lightweight" monitoring script.<br><br>On Friday, February 24, 2012 10=
:53:18 AM UTC-6, John Nunemaker wrote:<blockquote class=3D"gmail_quote" sty=
le=3D"margin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding-left:1e=
x">I installed mms on a slave database. From there it talks to both master/=
slave. We have pretty consistently been seeing the mms-agent grab almost al=
l the available memory. Probably happens like 5-10 times a day.<div><br></d=
iv><div><p style=3D"margin:0px 0px 0px 0px;font:12.0px Helvetica"><a href=
=3D"http://cl.ly/EV04" target=3D"_blank">http://cl.ly/EV04</a></p><p style=
=3D"margin:0px 0px 0px 0px;font:12.0px Helvetica"><br></p><p style=3D"margi=
n:0px 0px 0px 0px;font:12.0px Helvetica">Anyone else having issues with thi=
s? Anything else I can provide to help track this down?</p></div></blockquo=
te></blockquote></div></blockquote></div></blockquote></blockquote></div></=
blockquote></blockquote>
------=_Part_1104_17517867.1335204552279--

------=_Part_1103_31686791.1335204552278--
Create a group - Google Groups - Google Home - Terms of Service - Privacy Policy
Account Options
