[rbak-nsp] Error Log for PPA and QoS

[Tanan Satayapiwat]

There are quite a handful of bugs which lead to such the error message. Try upgrade to latest SEOS and see if you can get rid of the issue. 

6.1.3.9 would be a good candidate.

Tanan

-----Original Message-----
From: redback-nsp-bounces at puck.nether.net [mailto:redback-nsp-bounces at puck.nether.net] On Behalf Of Blake Willis
Sent: Wednesday, August 04, 2010 9:09 AM
To: redback-nsp at puck.nether.net
Subject: Re: [rbak-nsp] Error Log for PPA and QoS

Hi Alireza,

We've been hit by this recently; it started a few days after we made some changes to our clips subscriber QoS policy.  Interestingly our error is "already active" not :invalid parent id" so it may not be the same bug.  This is in 6.1.3.4.

When we see the error message it means that a subscriber has been authenticated & set up properly but they haven't been allocated a queuing point so the box can't forward packets to them.  This only happens to a small portion of subscribers, but it gets worse over time.  Rebooting makes the issue go away until the box runs out of unused queuing points & has to start recycling previously allocated ones (in our case with se100 with each subscriber using 1 queuing point this means that there have been 65k subscriber connections, which takes about a week).

The answer from the TAC (in our case) is that it's a known bug (basically it doesn't wait long enough for the newly allocated queuing  
point to be freed up) & that we should upgrade to 6.1.3.8 or later.   
If you upgrade, test again and report your findings to the TAC if you still see the bug.  We've planned this for some time after everyone gets back from summer holidays.

Your error is not exactly the same as ours, however. "invalid parent id" may mean that your box is trying to bind a queuing point to a node further down the chain that is no longer present.  I've seen this sort of thing happen when applying a complex QoS config all at once, or making a lot of changes to existing policies that are applied & in production.  My suggestion to you is to review your QoS configs from top to bottom & make sure that you understand the hierarchical  
relationship of each level of QoS and the     position in the tree  
that each port, circuit, subsciber, & node are bound to & make sure you haven't tried to do anything unsupported (e.g. using h-nodes on l2tp subscribers, etc).

  - have you made any changes to your qos configs before you saw this?

  - how long has it been since your last reboot?

  - how often do new subscribers connect to your box?

You may be able to get the message to go away by removing all statements that apply QoS policies to ports & subscribers ("qos rate", "qos policy" etc) and then putting them back in one at a time (commiting the config at each change and waiting a bit before applying the next one).

Good luck.  Let us know how it goes.

---
  Blake Willis
  Network Architect
  iBrowse

On 3 Aug 2010, at 19:00, redback-nsp-request at puck.nether.net wrote:

> From: "Alireza Soltanian" <soltanian at gmail.com>
>
> Hi
>
> Recently I received a huge amount of this log on my SE-100:
>
> Aug 3 12:44:12: %PPAQOS-3-TM_ERROR: f4156969/0000052988/680000000:02/
> EPPA/EU00:TM: failed to add L2 node 4511 to aggr node 0 - invalid 
> parent id - 0005cc04 00098618 0005c770 0002dd62 000267d4 000cca20 
> 00024a70 000a4372 00024650 0009511c 000950f4 00000000
>
> Any Idea what does it mean and How we can prevent it?
>
> Thanx
> Alireza
_______________________________________________
redback-nsp mailing list
redback-nsp at puck.nether.net
https://puck.nether.net/mailman/listinfo/redback-nsp