[j-nsp] JUNOS

keith tokash ktokash at hotmail.com
Tue Jun 23 14:07:40 EDT 2009


This thread made me curious, so I checked one of our M320s (on 8.3R3.4) and came up with a very similar memory utilization.

You're using ~90% of 768MB - 691MB
We're using ~30% of 2048MB - 634MB

So it looks like the memory requirements are creeping up if you're holding a full table (with its own creep...) and running newer code revs.  The inactive memory is a good point though, thanks.  Lurking for the win.



ktokash at br2.phx1-re1> show route all 

inet.0: 284591 destinations, 1056556 routes (284133 active, 0 holddown, 153288 hidden)


ktokash at br2.phx1-re1> show chassis routing-engine    
Routing Engine status:
  Slot 0:
    Current state                  Backup
    Election priority              Master
    Temperature                 32 degrees C / 89 degrees F
    CPU temperature             31 degrees C / 87 degrees F
    DRAM                      2048 MB
    Memory utilization          14 percent
<snip>
Routing Engine status:
  Slot 1:
    Current state                  Master
    Election priority              Backup
    Temperature                 34 degrees C / 93 degrees F
    CPU temperature             36 degrees C / 96 degrees F
    DRAM                      2048 MB
    Memory utilization          31 percent
<snip>


ktokash at br2.phx1-re1> show system processes brief 
last pid: 87827;  load averages:  0.10,  0.09,  0.03  up 557+13:25:17    18:00:38
63 processes:  1 running, 62 sleeping

Mem: 261M Active, 209M Inact, 123M Wired, 66M Cache, 69M Buf, 1347M Free
Swap: 2048M Total, 2048M Free



The information in this e-mail is intended for the 
attention and use of the everyone in the world, or I wouldn't have sent it in an unencrypted email. This message or any part thereof can and should be disclosed, copied, distributed and retained by any person without 
authorization from the addressee.  Furthermore, I reserve the right to disclose, copy, distribute and retain anything anyone sends *me* via email, up to and including putting the exact text in a MySpace bulletin.




> Date: Tue, 23 Jun 2009 13:29:01 -0400
> From: bmannella at teraswitch.com
> To: martin.mogensen at bt.com
> CC: juniper-nsp at puck.nether.net
> Subject: Re: [j-nsp] JUNOS
> 
> Ok heres what i got..
> 
> @ibr1.ash> show system processes brief
> last pid: 40751;  load averages:  0.00,  0.00,  0.00  up 211+21:54:08    13:47:2                                                                              8
> 112 processes: 3 running, 92 sleeping, 17 waiting
> 
> Mem: 405M Active, 136M Inact, 112M Wired, 57M Cache, 69M Buf, 32M Free
> Swap: 1536M Total, 1536M Free
> 
> so i guess i am not in all that bad of shape..
> 
> Can someone recommend the most stable version of Junos or the M series?
> 
> Brendan Mannella
> President and CEO
> TeraSwitch Networks Inc.
> Office: 412.224.4333 x303
> Toll-Free: 866.583.6338
> Mobile: 412-592-7848
> Efax: 412.202.7094
> 
> 
> 
> ----- Original Message -----
> From: "martin mogensen" <martin.mogensen at bt.com>
> To: bmannella at teraswitch.com
> Cc: juniper-nsp at puck.nether.net, sean at clarke-3.demon.nl
> Sent: Tuesday, June 23, 2009 11:49:56 AM GMT -05:00 US/Canada Eastern
> Subject: RE: [j-nsp] JUNOS
> 
> Brendan
> 
> You can try:
> show system processes brief 
> 
> to see how much memory is marked as inactive. The percentage calculation considers inactive memory as used memory which strictly speaking is correct (this is managed by the FreeBSD that JUNOS run on top of). However, inactive memory will be freed up automatically by the router if needed. This way the percentage value can look bad even though the router still has plenty of inactive memory left it can free up as needed. 
> 
> If the inactive memory is indeed high and you want to monitor the percentage, you can lauch a memory intentive task, ie compression/decompression of large files, which will free up inactive memory so the calculation will become more representative. Naturally you should only do this if there is indeed a large amount off inactive memory - you should not risk to run out of memory. 
> 
> Cheers
> Martin
> 
> -----Message d'origine-----
> De : juniper-nsp-bounces at puck.nether.net [mailto:juniper-nsp-bounces at puck.nether.net] De la part de Brendan Mannella
> Envoyé : mardi 23 juin 2009 16:58
> À : Sean Clarke
> Cc : juniper-nsp at puck.nether.net
> Objet : Re: [j-nsp] JUNOS
> 
> ibr1.ash> show route all
> 
> inet.0: 283638 destinations, 486468 routes (283619 active, 18 holddown, 2 hidden                                                                              )
> 
> ibr1.ash> show chassis routing-engine
> Routing Engine status:
>     Temperature                 30 degrees C / 86 degrees F
>     CPU temperature             28 degrees C / 82 degrees F
>     DRAM                       768 MB
>     Memory utilization          89 percent
>     CPU utilization:
>       User                       0 percent
>       Background                 0 percent
>       Kernel                     2 percent
>       Interrupt                  0 percent
>       Idle                      97 percent
>     Model                          RE-5.0
>     Serial ID                      
>     Start time                     2008-11-23 14:53:50 EST
>     Uptime                         211 days, 19 hours, 23 minutes, 19 seconds
>     Load averages:                 1 minute   5 minute  15 minute
>                                        0.05       0.10       0.04
> 
> I have a 2gig flash card installed, so is no issue.
> 
> I am just trying to figure out if its a software issue/bug causing this or its just the number of routes i have.
> 
> Brendan Mannella
> President and CEO
> TeraSwitch Networks Inc.
> Office: 412.224.4333 x303
> Toll-Free: 866.583.6338
> Mobile: 412-592-7848
> Efax: 412.202.7094
> 
> 
> 
> ----- Original Message -----
> From: "Sean Clarke" <sean at clarke-3.demon.nl>
> To: "Brendan Mannella" <bmannella at teraswitch.com>
> Sent: Tuesday, June 23, 2009 10:58:24 AM GMT -05:00 US/Canada Eastern
> Subject: Re: [j-nsp] JUNOS
> 
> Hi Brendan
> 
> How many routes etc do you have   ?
> The memory sounds a bit high utilised to me, if it stays at 91%.
> 
> There are many bug fixes from 9.1 to 9.5, of course, if you have a 1G flash card (or no flash card) then you can upgrade anyway ... the memory should not max out.
> 
> Typically the RE should be about 5% utilised, if the routes are stable
> 
> cheers
> 
> 
> 
> On 6/23/09 11:45 AM, Brendan Mannella wrote:
> > Hello,
> >
> > I have two M7i routers with RE-400-768s. One is running 9.1R1.8 and 
> > the other 8.3R4.3.
> >
> > They each have one transit link landed on them and very little IBGP/OSPF.
> >
> > The router with 9.1 on it has 91% memory usage while the 8.3 has 59%.
> >
> > Now I know the more recent code probably has more features and so 
> > forth. But I am worried that maybe that version of code has a memory 
> > leak or related issue.
> >
> > Does anyone know of any outstanding issues with 9.1R1.8?
> >
> > The router came with this version of code on it and I don¹t like the 
> > fact that its R1.8 but I am worried about upgrading to a more stable 
> > release as I don¹t want the memory maxed out.
> >
> > Any suggestions? What is the most stable code at this point for this 
> > platform?
> >
> > Thanks in Advance.
> >
> > Brendan
> > _______________________________________________
> > juniper-nsp mailing list juniper-nsp at puck.nether.net 
> > https://puck.nether.net/mailman/listinfo/juniper-nsp
> >
> >
> >    
> 
> _______________________________________________
> juniper-nsp mailing list juniper-nsp at puck.nether.net https://puck.nether.net/mailman/listinfo/juniper-nsp
> _______________________________________________
> juniper-nsp mailing list juniper-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/juniper-nsp

_________________________________________________________________
Insert movie times and more without leaving Hotmail®.
http://windowslive.com/Tutorial/Hotmail/QuickAdd?ocid=TXT_TAGLM_WL_HM_Tutorial_QuickAdd_062009


More information about the juniper-nsp mailing list