Mailing List Archive

This is on a 6.0R1.3 btw.
On Nov 10, 2003, at 7:58 PM, Phil Rosenthal wrote:

> Has anyone ever seen this error in /var/log/messages ?
>
> Nov 10 14:27:11 juniper /kernel: Nexthop index allocation failed:
> public index space exhausted
>
> I get that error about 600 times every 10 minutes.
>
> --Phil Rosenthal
> ISPrime, Inc.
>
> _______________________________________________
> juniper-nsp mailing list juniper-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/juniper-nsp
>
>
--Phil Rosenthal
ISPrime, Inc.

The kernal stores indexes to next-hops. There is only a limited number
of next-hop indexes that can be stored before memory is full and
produce the message you are seeing.

It would help to know your configuration. a 'show route summary' and
'show pfe route summary' just to get the idea of the size of your
routing tables.

Also what type of hardware you are using could help
Gary


On Monday, Nov 10, 2003, at 17:06 US/Pacific, Phil Rosenthal wrote:

> This is on a 6.0R1.3 btw.
> On Nov 10, 2003, at 7:58 PM, Phil Rosenthal wrote:
>
>> Has anyone ever seen this error in /var/log/messages ?
>>
>> Nov 10 14:27:11 juniper /kernel: Nexthop index allocation failed:
>> public index space exhausted
>>
>> I get that error about 600 times every 10 minutes.
>>
>> --Phil Rosenthal
>> ISPrime, Inc.
>>
>> _______________________________________________
>> juniper-nsp mailing list juniper-nsp@puck.nether.net
>> http://puck.nether.net/mailman/listinfo/juniper-nsp
>>
>>
> --Phil Rosenthal
> ISPrime, Inc.
>
> _______________________________________________
> juniper-nsp mailing list juniper-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/juniper-nsp
>

M20, dual SSB-E, dual RE

> show route summary

inet.0: 128188 destinations, 581434 routes (128094 active, 1 holddown,
127366 hidden)
Restart Complete
Direct: 54 routes, 54 active
Local: 53 routes, 53 active
BGP: 581265 routes, 127925 active
Static: 62 routes, 62 active

tablea.inet.0: 107 destinations, 107 routes (107 active, 0 holddown, 0
hidden)
Restart Complete
Direct: 54 routes, 54 active
Local: 53 routes, 53 active

tableb.inet.0: 107 destinations, 107 routes (107 active, 0 holddown, 0
hidden)
Restart Complete
Direct: 54 routes, 54 active
Local: 53 routes, 53 active

inet6.0: 484 destinations, 487 routes (484 active, 0 holddown, 0 hidden)
Restart Complete
Direct: 4 routes, 4 active
Local: 2 routes, 2 active
BGP: 479 routes, 476 active
Static: 2 routes, 2 active


> show pfe route summary


IPv4 Route Tables:
Index Routes Size(b)
-------- ---------- ----------
Default 128413 8595747
1 4 265
2 111 7538
3 111 7538

MPLS Route Tables:
Index Routes Size(b)
-------- ---------- ----------
Default 1 64

IPV6 Route Tables:
Index Routes Size(b)
-------- ---------- ----------
Default 495 33922
1 3 209


-P

On Nov 10, 2003, at 9:42 PM, Gary Tate wrote:

> The kernal stores indexes to next-hops. There is only a limited
> number of next-hop indexes that can be stored before memory is full
> and produce the message you are seeing.
>
> It would help to know your configuration. a 'show route summary' and
> 'show pfe route summary' just to get the idea of the size of your
> routing tables.
>
> Also what type of hardware you are using could help
> Gary
>
>
> On Monday, Nov 10, 2003, at 17:06 US/Pacific, Phil Rosenthal wrote:
>
>> This is on a 6.0R1.3 btw.
>> On Nov 10, 2003, at 7:58 PM, Phil Rosenthal wrote:
>>
>>> Has anyone ever seen this error in /var/log/messages ?
>>>
>>> Nov 10 14:27:11 juniper /kernel: Nexthop index allocation failed:
>>> public index space exhausted
>>>
>>> I get that error about 600 times every 10 minutes.
>>>
>>> --Phil Rosenthal
>>> ISPrime, Inc.
>>>
>>> _______________________________________________
>>> juniper-nsp mailing list juniper-nsp@puck.nether.net
>>> http://puck.nether.net/mailman/listinfo/juniper-nsp
>>>
>>>
>> --Phil Rosenthal
>> ISPrime, Inc.
>>
>> _______________________________________________
>> juniper-nsp mailing list juniper-nsp@puck.nether.net
>> http://puck.nether.net/mailman/listinfo/juniper-nsp
>>
>
>
--Phil Rosenthal
ISPrime, Inc.

This looks fine. I would open a case with the JTAC so that they can
look further into the issue with you.

Gary

On Monday, Nov 10, 2003, at 18:46 US/Pacific, Phil Rosenthal wrote:

> M20, dual SSB-E, dual RE
>
> > show route summary
>
> inet.0: 128188 destinations, 581434 routes (128094 active, 1 holddown,
> 127366 hidden)
> Restart Complete
> Direct: 54 routes, 54 active
> Local: 53 routes, 53 active
> BGP: 581265 routes, 127925 active
> Static: 62 routes, 62 active
>
> tablea.inet.0: 107 destinations, 107 routes (107 active, 0 holddown, 0
> hidden)
> Restart Complete
> Direct: 54 routes, 54 active
> Local: 53 routes, 53 active
>
> tableb.inet.0: 107 destinations, 107 routes (107 active, 0 holddown, 0
> hidden)
> Restart Complete
> Direct: 54 routes, 54 active
> Local: 53 routes, 53 active
>
> inet6.0: 484 destinations, 487 routes (484 active, 0 holddown, 0
> hidden)
> Restart Complete
> Direct: 4 routes, 4 active
> Local: 2 routes, 2 active
> BGP: 479 routes, 476 active
> Static: 2 routes, 2 active
>
>
> > show pfe route summary
>
>
> IPv4 Route Tables:
> Index Routes Size(b)
> -------- ---------- ----------
> Default 128413 8595747
> 1 4 265
> 2 111 7538
> 3 111 7538
>
> MPLS Route Tables:
> Index Routes Size(b)
> -------- ---------- ----------
> Default 1 64
>
> IPV6 Route Tables:
> Index Routes Size(b)
> -------- ---------- ----------
> Default 495 33922
> 1 3 209
>
>
> -P
>
> On Nov 10, 2003, at 9:42 PM, Gary Tate wrote:
>
>> The kernal stores indexes to next-hops. There is only a limited
>> number of next-hop indexes that can be stored before memory is full
>> and produce the message you are seeing.
>>
>> It would help to know your configuration. a 'show route summary' and
>> 'show pfe route summary' just to get the idea of the size of your
>> routing tables.
>>
>> Also what type of hardware you are using could help
>> Gary
>>
>>
>> On Monday, Nov 10, 2003, at 17:06 US/Pacific, Phil Rosenthal wrote:
>>
>>> This is on a 6.0R1.3 btw.
>>> On Nov 10, 2003, at 7:58 PM, Phil Rosenthal wrote:
>>>
>>>> Has anyone ever seen this error in /var/log/messages ?
>>>>
>>>> Nov 10 14:27:11 juniper /kernel: Nexthop index allocation failed:
>>>> public index space exhausted
>>>>
>>>> I get that error about 600 times every 10 minutes.
>>>>
>>>> --Phil Rosenthal
>>>> ISPrime, Inc.
>>>>
>>>> _______________________________________________
>>>> juniper-nsp mailing list juniper-nsp@puck.nether.net
>>>> http://puck.nether.net/mailman/listinfo/juniper-nsp
>>>>
>>>>
>>> --Phil Rosenthal
>>> ISPrime, Inc.
>>>
>>> _______________________________________________
>>> juniper-nsp mailing list juniper-nsp@puck.nether.net
>>> http://puck.nether.net/mailman/listinfo/juniper-nsp
>>>
>>
>>
> --Phil Rosenthal
> ISPrime, Inc.
>

On Nov 10, 2003, at 11:04 PM, Gary Tate wrote:

> This looks fine. I would open a case with the JTAC so that they can
> look further into the issue with you.
>
> Gary
I've already upgraded to 6.1R1.4, I will do it again if it reoccurs.

--Phil Rosenthal
ISPrime, Inc.

Phil,

an additional question.
is this a Route Reflector for vpn-routes ?
and perform the following.
show system virtual-memory | match rtn

Josef


Tuesday, November 11, 2003, 3:46:45 AM, you wrote:
> M20, dual SSB-E, dual RE

>> show route summary

> inet.0: 128188 destinations, 581434 routes (128094 active, 1 holddown,
> 127366 hidden)
> Restart Complete
> Direct: 54 routes, 54 active
> Local: 53 routes, 53 active
> BGP: 581265 routes, 127925 active
> Static: 62 routes, 62 active

> tablea.inet.0: 107 destinations, 107 routes (107 active, 0 holddown, 0
> hidden)
> Restart Complete
> Direct: 54 routes, 54 active
> Local: 53 routes, 53 active

> tableb.inet.0: 107 destinations, 107 routes (107 active, 0 holddown, 0
> hidden)
> Restart Complete
> Direct: 54 routes, 54 active
> Local: 53 routes, 53 active

> inet6.0: 484 destinations, 487 routes (484 active, 0 holddown, 0 hidden)
> Restart Complete
> Direct: 4 routes, 4 active
> Local: 2 routes, 2 active
> BGP: 479 routes, 476 active
> Static: 2 routes, 2 active


>> show pfe route summary


> IPv4 Route Tables:
> Index Routes Size(b)
> -------- ---------- ----------
> Default 128413 8595747
> 1 4 265
> 2 111 7538
> 3 111 7538

> MPLS Route Tables:
> Index Routes Size(b)
> -------- ---------- ----------
> Default 1 64

> IPV6 Route Tables:
> Index Routes Size(b)
> -------- ---------- ----------
> Default 495 33922
> 1 3 209


> -P

> On Nov 10, 2003, at 9:42 PM, Gary Tate wrote:

>> The kernal stores indexes to next-hops. There is only a limited
>> number of next-hop indexes that can be stored before memory is full
>> and produce the message you are seeing.
>>
>> It would help to know your configuration. a 'show route summary' and
>> 'show pfe route summary' just to get the idea of the size of your
>> routing tables.
>>
>> Also what type of hardware you are using could help
>> Gary
>>
>>
>> On Monday, Nov 10, 2003, at 17:06 US/Pacific, Phil Rosenthal wrote:
>>
>>> This is on a 6.0R1.3 btw.
>>> On Nov 10, 2003, at 7:58 PM, Phil Rosenthal wrote:
>>>
>>>> Has anyone ever seen this error in /var/log/messages ?
>>>>
>>>> Nov 10 14:27:11 juniper /kernel: Nexthop index allocation failed:
>>>> public index space exhausted
>>>>
>>>> I get that error about 600 times every 10 minutes.
>>>>
>>>> --Phil Rosenthal
>>>> ISPrime, Inc.
>>>>
>>>> _______________________________________________
>>>> juniper-nsp mailing list juniper-nsp@puck.nether.net
>>>> http://puck.nether.net/mailman/listinfo/juniper-nsp
>>>>
>>>>
>>> --Phil Rosenthal
>>> ISPrime, Inc.
>>>
>>> _______________________________________________
>>> juniper-nsp mailing list juniper-nsp@puck.nether.net
>>> http://puck.nether.net/mailman/listinfo/juniper-nsp
>>>
>>
>>
> --Phil Rosenthal
> ISPrime, Inc.

> _______________________________________________
> juniper-nsp mailing list juniper-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/juniper-nsp

On Nov 11, 2003, at 12:49 AM, Josef Buchsteiner wrote:

> Phil,
>
> an additional question.
> is this a Route Reflector for vpn-routes ?
> and perform the following.
> show system virtual-memory | match rtn
>
> Josef
Nope.

I already upgraded to 6.1R1.4, and rebooted, but here's what it says
now:
> show system virtual-memory |match rtn
rtnexthop, iffamily, temp, devbuf, zombie, proc-args, dev_t,
ipfw, rtnexthop, iflogical, iftable, temp, devbuf, proc-args,
rtnexthop 510 100K 211K127940K 221948 0 0 128,256

--Phil Rosenthal
ISPrime, Inc.