Catalyst 3750 Ingress SPQ/SRR behavior

Do Cisco engineers review this community at all?
I am working on the latest version of QoS standard for our Enterprise and noticed the following conflicting information officially provided by Cisco.
My question relates to ingress/pre-ring Strict Priority Queue (SPQ) logic.
Cisco Catalyst 3750 QoS Configuration Examples document states that SPQ on ingress is configured and serviced as follows
mls qos srr-queue input priority-queue 2 bandwidth 10
mls qos srr-queue input bandwidth 90 10
SPQ services Q2 up to the configured 10% of ingress bandwidth
Any excessive traffic in Q2 is not dropped, but is serviced by SRR in accordance with the configured weights
For example, a momentary 5Gbps of aggregated ingress EF traffic will be serviced in the following way
SQP services 10% of total ring's bandwidth, or 3.2Gbp, leaving 1.8Gbps for SRR processing
SRR services excessive 1.8Gbps in accordance w/ weights Q1 - 90 and Q2 - 10, such as Q1 gets 25.92Gbps and Q2 get 2.88Gbps more.
The following pictures provides in-depth look into Ingress queuing logic.
Alternatively, Cisco Medianet Campus Design v4.0 provides the following example w/ comments
C3750-E(config)#mls qos srr-queue input priority-queue 2 bandwidth 30
! Q2 is enabled as a strict-priority ingress queue with 30% BW
C3750-E(config)#mls qos srr-queue input bandwidth 70 30
! Q1 is assigned 70% BW via SRR shared weights
! Q2 SRR shared weight is ignored (as it has been configured as a PQ)
Basically, they now say Q2 bandwidth weight is ignore because it is configured as Strict Priority Queue. Doesn't it look contradictory?
In my humble opinion Medianet (or SRND v4.0!!!) provides an incorrect information re ingress queuing on Catalyst 3750 platform.
I am not sure I can easily test it, providing that an internal ring must experience a congestion. I don't think I can send more than 32Gbps of traffic into any of my lab 3750 switches.
Also, I don't think this mistake can be critical in my environment as I don't expect to have momentary full capacity load on those... but it can be critical for others.
Much appreciate
Tim

Similar Messages

Policer with IPv6 class-map on Catalyst 3750

Hi,
I've the following problem.
It's my goal to ratelimit incoming IPv6 traffic dependent on the destination IP address range.
On a Catalyst 3750 (Image: c3750-ipservicesk9-mz.122-55.SE1.bin) I've set up the configuration as follows:
mls qos
ipv6 access-list DESTINATION-RANGE-A
permit ipv6 any 2007::/16
ipv6 access-list DESTINATION-RANGE-B
permit ipv6 any 2B03::/16
class-map match-all A
match access-group name DESTINATION-RANGE-A
class-map match-all B
match access-group name DESTINATION-RANGE-B
policy-map RL-POLICY
class A
police 2000000 8000 exceed-action drop
class B
police 6000000 8000 exceed-action drop
interface GigabitEthernet1/0/7
switchport access vlan 90
load-interval 30
service-policy input RL-POLICY
The last CLI command which should bind the policy to the specific interface, leads to the following error message
QoS: class(A) IPv6 class not supported on interface GigabitEthernet1/0/7
Are hardware/software limitations the reason for this behavior or is there any misconfiguration?
Thanks in advance for your help!
Regards,
Jens

If you are thinking of IPv6 prefix I tried everything. From /128 for single host to /64, nothing works.

Ratelimiter with IPv6 class-map on Catalyst 3750

Hi,
I've the following problem.
It's my goal to ratelimit incoming IPv6 traffic dependent on the destination IP address range.
On a Catalyst 3750 (Image: c3750-ipservicesk9-mz.122-55.SE1.bin) I've set up the configuration as follows:
mls qos
ipv6 access-list DESTINATION-RANGE-A
permit ipv6 any 2007::/16
ipv6 access-list DESTINATION-RANGE-B
permit ipv6 any 2B03::/16
class-map match-all A
match access-group name DESTINATION-RANGE-A
class-map match-all B
match access-group name DESTINATION-RANGE-B
policy-map RL-POLICY
class A
police 2000000 8000 exceed-action drop
class B
police 6000000 8000 exceed-action drop
interface GigabitEthernet1/0/7
switchport access vlan 90
load-interval 30
service-policy input RL-POLICY
The last CLI command which should bind the policy to the specific interface, leads to the following error message
QoS: class(A) IPv6 class not supported on interface GigabitEthernet1/0/7
Are hardware/software limitations the reason for this behavior or is there any misconfiguration?
Thanks in advance for your help!
Regards,
Jens

If you are thinking of IPv6 prefix I tried everything. From /128 for single host to /64, nothing works.

Password problem on Catalyst 3750

Hi all,
I am an Unix administrator and we lost all account and password informations to connect on a Catalyst 3750 switch.
Is there a way to connect to the switch (With a serial cable?) an to create a new account without losing the configuration ?
Thank's for your reply.
Regards.

This method won't reset the configuration.
During the boot process, you rename the current configuration so it doesn't get loaded.
rename flash:config.text flash:config.old
When the switch boots, it loads a blank image. Then you rename the config file and load it into the running config and you can change the logon credentials to something you know.
It's important that you follow the steps exactly so you don't lose the configuration. Password recovery is a pretty common procedure.
The only other method of recovery is if you have a copy of the current running configuration. Send me a private message if you have this config and we can discuss the possibility.

Catalyst (3750 24 10/100/1000T) and (3750 12 SFP) Stacking Problems

Dear all
I'm having a very strange situation here (at least for me)
we have 4 core switches
2 x   WS-C3750G-24T-S Catalyst 3750 24 10/100/1000T + IPB Image
and
2 x   WS-C3750G-12S-S Catalyst 3750 12 SFP + IPB Image
Stack configuration is done this way
when the switches are powered on, the first two SFP core switches are seen as a single stack with the stack master LED turned green on the first switch
the other two (24 10/100/1000T) switches have the RPS LEDs always green, mode cannot be changed, and cannot be accessed by Console connection
but when the (24 10/100/1000T) are powered off, the first (SFP) switch in the stack reports that " Switch 3 and 4 has been removed from Stack "
which means they are stacked but there's something wrong, because
only the SFP ports are shown in the " Show interfaces status " , the ethernet ports of the bottom switches are not present !!!
can you please tell me what's the poblem ?

Dear Daniel
Sorry for my delayed response but i was actually quite busy
but the problem was actually in another sense
the default profile for the Catalyst 3750 SFP is the Aggregate SDM Template
while the 3750 10/100/1000 ethernet Switch Default SDM profile was Desktop profile
so i had an SDM mismatch
DATACENTER#sh switch detail
Switch/Stack Mac Address : 081f.f3cf.1c80
                                           H/W   Current
Switch# Role   Mac Address     Priority Version State
*1       Master 081f.f3cf.1c80     1      0       Ready
2       Member 081f.f3cf.5900     1      0       Ready
3       Member aca0.16ac.0180     1      2       SDM Mismatch
4       Member aca0.16a3.bc80     1      2       SDM Mismatch
         Stack Port Status             Neighbors
Switch# Port 1     Port 2           Port 1   Port 2
1        Ok         Ok                2        4
2        Ok         Ok                3        1
3        Ok         Ok                4        2
4        Ok         Ok                1        3
all i did was changing the default profile of the SFP switches into the Desktop Profile and problem was solved
switch 1 provision ws-c3750g-12s
switch 2 provision ws-c3750g-12s
switch 3 provision ws-c3750g-24t
switch 4 provision ws-c3750g-24t
system mtu routing 1500
ip subnet-zero
no file verify auto
spanning-tree mode pvst
spanning-tree extend system-id
vlan internal allocation policy ascending
interface GigabitEthernet1/0/1
interface GigabitEthernet1/0/2
interface GigabitEthernet1/0/3
interface GigabitEthernet1/0/4
interface GigabitEthernet1/0/5
interface GigabitEthernet1/0/6
interface GigabitEthernet1/0/7
interface GigabitEthernet1/0/8
interface GigabitEthernet1/0/9
interface GigabitEthernet1/0/10
interface GigabitEthernet1/0/11
interface GigabitEthernet1/0/12
interface GigabitEthernet2/0/1
interface GigabitEthernet2/0/2
interface GigabitEthernet2/0/3
switchport trunk encapsulation dot1q
switchport mode trunk
interface GigabitEthernet2/0/4
interface GigabitEthernet2/0/5
interface GigabitEthernet2/0/6
interface GigabitEthernet2/0/7
interface GigabitEthernet2/0/8
interface GigabitEthernet2/0/9
interface GigabitEthernet2/0/10
interface GigabitEthernet2/0/11
interface GigabitEthernet2/0/12
interface GigabitEthernet3/0/1
interface GigabitEthernet3/0/2
interface GigabitEthernet3/0/3
interface GigabitEthernet3/0/4
interface GigabitEthernet3/0/5
interface GigabitEthernet3/0/6
interface GigabitEthernet3/0/7
interface GigabitEthernet3/0/8
interface GigabitEthernet3/0/9
interface GigabitEthernet3/0/10
interface GigabitEthernet3/0/11
interface GigabitEthernet3/0/12
interface GigabitEthernet3/0/13
interface GigabitEthernet3/0/14
interface GigabitEthernet3/0/15
interface GigabitEthernet3/0/16
interface GigabitEthernet3/0/17
interface GigabitEthernet3/0/18
interface GigabitEthernet3/0/19
interface GigabitEthernet3/0/20
interface GigabitEthernet3/0/21
interface GigabitEthernet3/0/22
interface GigabitEthernet3/0/23
interface GigabitEthernet3/0/24
interface GigabitEthernet4/0/1
interface GigabitEthernet4/0/2
interface GigabitEthernet4/0/3
interface GigabitEthernet4/0/4
interface GigabitEthernet4/0/5
interface GigabitEthernet4/0/6
interface GigabitEthernet4/0/7
interface GigabitEthernet4/0/8
interface GigabitEthernet4/0/9
interface GigabitEthernet4/0/10
interface GigabitEthernet4/0/11
interface GigabitEthernet4/0/12
interface GigabitEthernet4/0/13
interface GigabitEthernet4/0/14
interface GigabitEthernet4/0/15
interface GigabitEthernet4/0/16
interface GigabitEthernet4/0/17
interface GigabitEthernet4/0/18
interface GigabitEthernet4/0/19
interface GigabitEthernet4/0/20
interface GigabitEthernet4/0/21
interface GigabitEthernet4/0/22
interface GigabitEthernet4/0/23
interface GigabitEthernet4/0/24
that's it !
: D
cheers

Debian Linux Bonding and Cisco Catalyst 3750 - best practise?

Hello everybody,
I would like to know what's best practice to do this:
The two NICs of a Debian Linux server wants to be connected with two Switchports of a Cisco Catalyst 3750 switch(stack). My goal is to have load-balancing and failover.
My /etc/network/interfaces looks like this:
iface bond0 inet static
       address 192.168.0.30
       netmask 255.255.255.0
       network 192.168.0.0
       broadcast 192.168.0.255
       gateway 192.168.0.1
       dns-nameservers 192.168.0.10 192.168.0.20
       dns-search xyz.mycompany.com
       slaves eth0 eth1
       bond_mode ???
       bond_miimon 100
       bond_downdelay 200
       bond-updelay 200
First question: What bond mode should I use?
The switchports looks like this:
interface GigabitEthernet3/0/4
switchport access vlan 20
switchport mode access
spanning-tree portfast
What changes are necessery here? Something like this?
interface GigabitEthernet3/0/4
switchport trunk encapsulation dot1q
switchport mode trunk
spanning-tree portfast
Thanks a lot for suggestions, hints, etc.! :-)
Greets
Stephan

Hi Michael,
thanks a lot for your answer - and sorry for my late reply!
I like to show you my solution - I hope that it is a solution. ;-)
My config on the switch(stack):
switch#show etherchannel summary
Group Port-channel Protocol    Ports
------+-------------+-----------+-----------------------------------------------
2      Po2(SU)         LACP      Gi3/0/3(P) Gi4/0/3(P)
switch#show running-config interface GigabitEthernet 3/0/3
Building configuration...
Current configuration : 172 bytes
interface GigabitEthernet3/0/3
description myserver, eth0
switchport access vlan 20
switchport mode access
channel-group 2 mode active
spanning-tree portfast
end
lansw01#show running-config interface GigabitEthernet 4/0/3
Building configuration...
Current configuration : 172 bytes
interface GigabitEthernet4/0/3
description myserver, eth1
switchport access vlan 20
switchport mode access
channel-group 2 mode active
spanning-tree portfast
end
switch#show running-config interface port-channel 2
Building configuration...
Current configuration : 82 bytes
interface Port-channel2
switchport access vlan 20
switchport mode access
end
The /etc/network/interfaces of my Debian machine looks like this:
auto lo
iface lo inet loopback
auto bond0
        iface bond0 inet static
        address 192.168.1.xxx
        netmask 255.255.255.0
        gateway 192.168.1.xxx
        dns-nameservers 192.168.1.xxx
        dns-search xxx.xxx.xxx
        bond-mode 4
        bond-miimon 100
        bond-downdelay 200
        bond-updelay 200
        bond-lacp-rate 1
        slaves eth0 eth1
This setup seems to work well. But I'm wondering that there is nothing with "trunking" in my setup. Would you like to give me your opinion about this?
Thanks a lot and many greets
Stephan

Make Fiber Ports Live on Catalyst 3750?

Hello All,
Have 2 Catalyst 3750's connected via fiber to the cisco gbics converts. Both gbic's are plugged into Slot 1 of 4 on the switches. Everything is physically hooked up correctly. My question is what command do I type to make the fiber ports active? I know the fiber led should go orange for 30 sec then green when you plug the fiber cable in the gbic but it's not working because I've got to make that port or interface active first right? Any help would be appreciated. Thanks in advanced!

Double check the cabling. Unplug one of the patch cables from one of the gbics. Look at which connector on the gbic has the light coming out of it. Compare to the fiber patch for which connector has light coming out of it. Ensure you have the fiber plugged into the gbic such that light from one connects to dark on the other. Do this at both ends.
Also, pull the gbics out of the switches & validate both are appropriate for the fiber type you are using. 1000-LX is for single mode fiber; 1000-SX is for multimode fiber.
If you have the interface administratively shut down, this will fix it:
int gig 0/1
no shut

Interconnecting Catalyst 3750 and 2948G-L3

I am trying to interconnect a Catalyst 3750 and a 2948G-L3 using fiber GBIC. The interfaces where the GBIC and fiber are attached show up as physically down. I have tried different ports and also changed both switches. No Luck. If I connect a 3524 to the 3750 using the same connection it works.
Are 2948G-L3 switches compatible with the 3750?
Thanks,
VT

Should have no problem. Can you try the following on the 3750's gig interface:
speed nonegotiate
See of the link comes up.
Please rate all posts.

Could connect two Nexus 5648 to a stack of Catalyst 3750?

Good Morning,
Can I connect a two Nexus 5600 (switchs distribution) with switch Catalyst 3750 (switch access)?
I have tried to integrate the Nexus switch in the network as switch distribution, replacing a Catalyst 6500 switch and does not work. Everything is going very slow. Although all the ping function properly.
Is there any incompatibility connect two nexus 5648 to a stack of Catalyst 3750 by VPC?
Thank you.
Greetings.

Hi,
There can be many reasons for the switches not joining the stack like SDM template mismatch.
You may go through the following document:
http://www.cisco.com/c/en/us/td/docs/switches/lan/catalyst3750/software/troubleshooting/switch_stacks.html#wp40112
Thanks
Ankur
"Please rate the post if found useful"

Catalyst 3750 Switch

How many total Vlans can you create in Catalyst 3750 Switch ? I read a document about Catalyst 3750 Switch. This document is said that "Although the switch stack supports a total of 1005 (normal-range and extended-range) VLANs, the
number of routed ports". However, I am not sure. Can you confirm for me ? Thanks

this link should be of some help to answer your question.
http://www.cisco.com/en/US/products/hw/switches/ps5532/products_command_reference_chapter09186a00803ec324.html#wp1031710
HTH-Cheers,
Swaroop

Revised Visio stencils for CRS and Catalyst 3750

We just submitted the CRS and Catalyst 3750 revisions to the Cisco web team for posting. They should be available within a day or so.
Regards,
Brett Newman
Cisco Visio Development
Visimation Inc.

Hi Kevin,
We updated the 3750 on 2/23/12 and the CRS on 2/20/12. Please check the download page.
Regards,
Brett Newman
Cisco Visio Development
Visimation Inc.

System FCS error frames on Catalyst 3750

Please,
may someone explain what are the "System FCS error frames" I see with "show controllers ethernet-controller" on some ports of a Catalyst (on trunk and also on access ports) WS-C3750G-24TS-S (3 switches in stackwyse)?
Error counters advance slowly, but they do grow... and I am afraid this could be the reason for variable slow access time to a server application menu...
For example "8 System FCS error frames".
The ports are connected to a PIX-515E and I see the same error count on the firewall interfaces (PIX "show interface" command shows me "8 input errors, 8 CRC, 0 frame, 0 overrun, 8 ignored, 0 abort")
A few errors may be seen also on other ports, connected to other (C3550)switches.
I tried to upgrade the IOS version, so I loaded "ipbase 12.2(25)SEB4 " on the Catalyst 3750, but it did not solve the problem.
PIX firewalls are running software 6.3(4).
Why these errors are called "System FCS error frames" ? I could not find anything searching on Cisco CCO!
Thanks

I apologies for the delay answering, my notebook has been down for 5 days and also cco access password...
Thank you for the explanation and suggestion.
We have a second rack with another three switch 4750 in stackwyse, same models as the "faulty" ones, and the errors are not present on those switches, so it must be something hardware related on the first group of switches.
I cannot touch those devices at the moment, we captured traffic frames with 3 Notebook PCs running "Ethereral" lan sniffer, traffic to and from the server and two clients and we discovered that in different ports of those switches, some packets never reach the clients, causing many TCP retransmissions, especially witch continuous traffic (http get and ftp transfers). This problem is visible on some, but not all, of the catalsyt 3750 switches in stackwyse, but in every one of the three switches. We are planning to go on-site and replace the hopefully faulty one ( or reseat the cables as you suggested).
Is this a problem you have seen before?
Regards
Franco Feri

New Visio Stencils for Catalyst 3750, Telepresence 1300

In response to customer requests, there are new images on the Visio Stencils web page at http://www.cisco.com/en/US/products/hw/prod_cat_visios.html
The updates include images for:
ASA 5510 Rear
ASA 5520 Front
ASA 5520 Rear
ASA 5540 Front
ASA 5540 Rear
ASA 5580 Front
ASA 5580 Rear
ME-C6524GT-8S Front
ME-C6524GT-8S Rear
ME-C6524GS-8S Front
ME-C6524GS-8S Rear
PWR-400W-AC - 400W AC
PWR-400W-DC - 400W DC
Catalyst 3750 Metro Front
Catalyst 3750 Metro Rear
PWR-ME3750-AC
PWR-ME3750-DC
CTS-1300-47
CTS-1300-65

Hi Kevin,
We updated the 3750 on 2/23/12 and the CRS on 2/20/12. Please check the download page.
Regards,
Brett Newman
Cisco Visio Development
Visimation Inc.

Trunking Catalyst 2950 to Catalyst 3750 problem

I cannot seem to figure out how to trunk a catalyst 3750 to a Catalyst 2950.
I've set
3750(config)#interface fastethernet 1/0/2
switchport mode trunk
switchport trunk encapsulation dot1q
BUT my Catalyst 2950 does not offer the "switchport trunk encapsulation dot1q" command
My 2950 is running IOS version 12.1(20)EA1a
is there a work around for this situation. Our network still employs a bunch of these 2950's.

The Catalyst 2950 series can only do 802.1Q trunking. It's the default, and only, choice. So there's no need to specify it when trunking.
In fact, since you don't have a choice of which encapsulation to use, there's no need for a "switchport trunk encapsulation" command. Which is why it's missing from the Cat2950 switch IOS.
This took me by surprise too, when I first transitioned out of the 3500XL series into 2950 and 3550 switches.

/31 on Catalyst 3750 routed port

I've discovered the feature 'Using 31-Bit Prefixes on IPv4 Point-to-Point Links' and I'm interested in implementing it on a point-to-point uplink network that terminates on one end to a Catalyst 3750 routed FastEthernet port ("no switchport").
I can enter the IP address in the interface config (172.16.100.0 255.255.255.254), and I get the warning "% Warning: use /31 mask on non point-to-point interface cautiously". Since I can guarantee that this is in fact a point-to-point uplink to a Layer 3 device, can I safely ignore the warning? Is there any way to configure the interface explicitly as a Point-to-Point interface instead of a broadcast interface?
Thanks!
-Mason

Hi,
It doesn't appear the error is platform specific. I am getting the same error in my lab 2600 when configuring a /31 bit mask.
You would use the /31 bit mask only on a p-t-p interface hence, you don't need to tell the interface it's a p-t-p interface. I don't know of any generic command to tell the broadcast interface that it's a p-t-p interface. There is a OSPF command but that doesn't have anything to do with this setup.
R2(config)#int e0/0
R2(config-if)#ip add 1.1.1.0 255.255.255.254
% Warning: use /31 mask on non point-to-point interface cautiously
R2(config-if)#^Z
R2#show run
*Mar 1 18:20:21.795: %SYS-5-CONFIG_I: Configured from console by console int e
/0
Building configuration...
Current configuration : 78 bytes
interface Ethernet0/0
ip address 1.1.1.0 255.255.255.254
half-duplex
end
R2#show ip arp
Protocol Address Age (min) Hardware Addr Type Interface
Internet 1.1.1.0 - 0004.dd85.eee0 ARPA Ethernet0/0
R2#ping 1.1.1.1
Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 1.1.1.1, timeout is 2 seconds:
Success rate is 80 percent (4/5), round-trip min/avg/max = 1/3/4 ms
R2#show ip arp
Protocol Address Age (min) Hardware Addr Type Interface
Internet 1.1.1.1 0 0004.dd85.e9c0 ARPA Ethernet0/0
Internet 1.1.1.0 - 0004.dd85.eee0 ARPA Ethernet0/0
HTH
Sundar

Catalyst 3750 Ingress SPQ/SRR behavior

Similar Messages

Maybe you are looking for