Cisco Switching/Routing :: 3750 Priority Queue Out Transmit Discards
Dec 13, 2011
After opening up Solarwinds NPM, I noticed that a few of my interfaces had lots of discards (who knows how long it's been sets the counters were reset)
interface GigabitEthernet1/0/25description Etherchannel to MamaCassswitchport trunk encapsulation dot1qswitchport mode trunkswitchport nonegotiatepriority-queue outchannel-group 4 mode on
interface GigabitEthernet2/0/25description Etherchannel to MamaCassswitchport trunk encapsulation dot1qswitchport mode trunkswitchport nonegotiatepriority-queue outchannel-group 4 mode on
interface Port-channel4switchport trunk encapsulation dot1qswitchport mode trunkswitchport nonegotiate,It looks as if priority-queue was configured outbound on these interfaces, could this be the cause of the transmit discards which are now up to 79,835, I just reset the counters on the interfaces a little while ago.
I'm not the best in the world when it comes to QoS, we do have some VoIP phones, but they are only a specific network, and do not travel outside, since there are used mainly for VoIP training. I do know both interfaces are running the default of FIFO.
We have a Cisco 6513 Switch. During some parts of the day it is having large amounts of transmit discards on a certain interface because of the amount of data we are pushing. We have determined it is a link problem and not a problem with the switch. The link cannot handle the amount of packets the switch is pushing out. I was wondering if there is a way to create a buffer or a way to regulate the maximum amount of packets that can leave that interface without losing any packets?
I have a Cisco Catalyst 2960 with IOS Release12.2(53)SE (because of a contract I can not update it) -> the release notes for this version describe the following:
When auto-QoS is enabled on the switch, priority queuing is not enabled. Instead, the switch uses shaped round robin (SRR) as the queuing mechanism. The auto-QoS feature is designed on each platform based on the feature set and hardware limitations, and the queuing mechanism supported on each platform might be different. There is no workaround. (CSCee22591)
My config is as follows:
interface FastEthernet0/1 switchport access vlan 200 switchport mode access srr-queue bandwidth share 10 10 60 20 priority-queue out mls qos trust dscp auto qos voip trust no cdp enable network-policy 1 spanning-tree portfastMy question now is:When the priority queue is not enabled with auto-qos because of the software bug is it nevertheless enabled with the additional priority-queue out command?
i have a 3560 connecting to a sp with limited bandwidth. i have one interface on the switch whose traffic i do not want to drop. i want this traffic to go into the high priority queue. i am not sure how this should be configured, but here is my best guess and my current qos configuration on the switch:
I am trying to implement priority queuing (LLQ) on a pair of 10GE links between a 4507 with Sup6E and a 4948 which are configured as an etherchannel. I am unable to configure a priority queue on the 4507. I am running into the following issues:
I want to have a priority queue for voice traffic and specify minimum bandwidth for a critical application. If I configure a class with the priority command it will not let me use the bandwidth command on another class unless the priority class is policed. If I try it without the police command I get the message "bandwidth kbps/percent command cannot co-exist with strict priority in the same policy-map ". If I add a police statement to the priority class then I don't get this error.
When I try to apply the resulting service-policy to the physical interface it says "% A service-policy with non-queuing actions should be attached to the port-channel associated with this physical port" and does not add the command to the config.
If I try to associate the same policy-map to the port-channel rather than the physical interface it says "% A service-policy with queuing actions can be attached in output direction only on physical ports" and does not add the command to the config.
All of the other interfaces on the 4500 are working OK. The trunks have auto qos voip trust configured and access ports are marking the critical application traffic.
The 4507 is running 12.2(44)SG1 EnterpriseK9. I don't have the luxury to upgrade blindly to fix the problem unless I can identify a specific bug that is causing the problem.
I have an issue where I'm seeing output discards on pretty much all my ports configured for QoS. The switches are cisco WS-C3750V2-48PS running 12.2(50)SE1. There are four switches stacked using stack cables. The QoS implemeted was auto-qos with no modifications to the standard config. All ports are in queue-set 1. The phones connected are Cisco 7942's. Already did the standard check for speed duplex mismatch, crc's, runts, giants, etc...No discards before the QoS was applied. No bug ID's I could find regarding these switches and this IOS version. The one thing noticed is that 99% of all the drops are from queue's 2 and 4 or 1 and 3 doing the below command.
**I've limited the cut and paste as to not clutter the discussion until someone requests something else**
show platform port-asic stats drop Port 18 TxQueue Drop Statistics Queue 0 Weight 0 Frames 0 Weight 1 Frames 0 [Code]....
I have an AP1242 near a couple of conference rooms. I get complaints when we have a large contingent in there and some try to download documents. The bandwidth consumption is meage - less than 2Mbps. But there could be say 50 people in those training rooms. I see no errors at the FE interface on the AP but I do see a good number of Transmit Discards.
Are 50 users connected to an AP1242 simply more than should be expected of the device? 802.11g is the radio.
Are transmit discards indicative of some kind of configuration problem? The users in these cases are generally very close to the access point - within a few feet even.
If it is indicative of a configuration problem would that be an issue of the configuration of the laptop wifi cards or a problem at the AP?
I am working on a QoS design which I hope to test at some point, but at this stage its from the books.My question is how to decide which queue and threshold to use for video traffic, then lower priority traffic.I understand the shaping and sharing commands, its the queuing and threshold bit I'm not clear on.The plan is to use the priority-queue for EF marked voice, this will be policed on ingress to provide an upper limit to EF traffic levels, then my second priority traffic will be video. Which queue will get serviced first once the priority queue is empty, and how do I decide which threshold to allocate my video traffic to? The document ion is not at all clear, I want to prioritse my traffic in the following order:
1 voice, use the priority queue 2 video, this to get serviced ahead of data, after voice. 3 interactive data 4 Bulk data 5 Best effort
So Q1 settings are ignored due to priority queue. Q2 gets 70%, Q3 25% etc.Is it as simple as putting video into Q2 T1, then interactive data into Q2 T2, will Q2T1 get a higher priority over Q2 T2 once the PQ is serviced?
We currently have a site with a very simple topology that uses a 3750X switch stack for a collapsed core. Everyday, the users have a conference call and experience poor voice quality.Its not bad when users call from several conference phones, but when everyone calls in on individual phones, there is choppy and almost inaudible voice quality experienced. The voice traffic flow would be as follows: Phone <-> 3750 switch <-> Voice GW We have packet captures showing that RTP packet loss is occuring from the phone to the voice gateway, but none from the voice gateway to the phones. We also have drops in the output queues that match drops on the asics. I can reset the counters and they will be clear until the call, and then they increment significantly during the call. The voice gateway and phones are non-Cisco. The switch stack has 6 switches. We are trusting the DSCP settings on the phones. All the queue drops from the phones are usually in queues 0-3, but all drops on the voice gateway is in queue 0. Below are the QoS settings; they are mostly default and we have not changed any queuing, thresholds, or buffers. Should we specify larger buffers and threshold for a designated queue and send EF traffic to that queue?
MySwitch#sh mls qos QoS is enabled QoS ip packet dscp rewrite is disabled Typical Port GigabitEthernet1/0/4 trust state: trust dscp
I've been working on breaking down and understanding the default auto qos configuration on a Cisco 3750 in the hopes of putting together a QoS strategy that will fit our environment. I'm having some difficulty understanding how the "mls qos queue-set output" syntax works.
From another post, at [URL], the author offers the following example and explanation;
How come there is syntax stating "threshold 2" when in the succeeding part the 400 refers to thresshold 1 and threshold 2 again? The syntax 400 400 is, apparently, already referring to thresshold 1 and 2, no?
I feel that 3560 and 3750 perform differently with the following two commands:
srr-queue bandwidth shape 5 0 0 0 srr-queue bandwidth limit 50 On 3750, the bandwidth for queue 1 is limited to 100mbps x 50% / 5 = 10mbps On 3560, the bandwidth for queue 1 is limited to the smaller value of BW / shape weight and BW x limit%.
Does it sound about right? is there a way to check for mls qos input queue drops? The show mls qos interface xxx stat only shows the output queue drops. Maybe for some reason the input queue never drops?
When configuring QoS on 3750s/3560s, we're mapping packets to particular interface output queues with commands such as: [code] The command to see what's actually being enqueued, dropped, etc. is: [code]
Note that these queues are numbered 0 - 3, and not 1 - 4. We've been assuming that the first queue number in the "mls qos" (i.e., 1) command maps to the first queue (i.e., 0) in the "show mls qos" command.
I've been fighting what seems to be an increased number of outqueue drops on our core stack and edge switches for the last 3 or 4 weeks.(The core consists of a stack of 5 3750s in 32-gig stack mode. The wkgrp switches are 3560s. all are at 12.2.52) The wkgrp switches are directly connected to users. We use Nortel IP phones with the phone inline with the user PC. auto-neg to 100/full. [code] However I have tried turning off QOS on a couple of workgroup switches (no mls qos, but left individual port configurations the same) but am still seeing drops.Since I have disabled qos on the switches in question (no mls qos) (not the core tho) I am presuming these commands have no affect on the switch operation and therefore cannot be related to the problem. With QOS turned off one would presume that it is general congestion - especially at the user edge where busy PC issues might contribute. So I wanted to see if I could see any instances of packets in the output queues building up.
I wrote some scripts and macros that essentially did a snapshot of 'show int' every 20 seconds or so, and looked for instances of 'Queue: x/' where x was greater than zero.What I found after several days of watching the core stack, and a few of the workgroup switches that are most often displaying the behavior, was that I NEVER saw ANY packets in output queues. I often saw packets in Input queues for VLAN1, once in a great while I would see packets on input queues for fa or Gi interfaces, but NEVER on output queues. [ code] Additionally, when I look (via snmp) at interface utilization on interfaces showing queue drops (both core and wkgroup), they are occurring at ridiculously low utilization levels (as low as 4 to 8%). I've tried to look for microbursts between the core and a wkgroup switch where the core interface was experiencing drops, but haven't seen any (using observer suite). [code] While the queue-drop counts aren't critically high at this point, they are happening more frequently than in the past and I would like to understand what is going on... In most cases, no error counters are incrementing for these interfaces. Is there some mechanism besides congestion that could cause output queue drops?
I have a AIR-AP1121G-A-K9 running c1100-k9w7-tar.123-7.JA2 (Autonomous)We have monitoring setup with Orion NPM and we consistently see output errors, Transmit discards and big buffer errors The users at the site have not reporting any issues but was wondering how to prevent these or are these normal?What causes the output errors on Wireless Radio ? How to troubleshoot further ?
Radio0-802.11G Total Output Errors 0 47749 Small Buffer Misses 4 misses 139 misses
We're testing the reference system shown in the figure below. System Description Four 2960 switches are used for transport;Equipment 1 and Equipment 2 exchange packets for synchronization;To reach synchronization Equipment 1 and 2 must exchange data with a very low jitter. 2960 Configuration details Four our test puprose, we're using 100Mbit/s ports (22 and 23) as trunk.In order to obtain minimum jitter We performed these configurations:We Enabled QoS;We Marked Synchronization packets with CoS 7 and DSCP 63;We marked other kind of traffic inserted in different ports) with CoS 0;We set "trust DSCP" on trunk ports;On the trunk ports we mapped traffic with CoS 7/DSCP 63 (and only this) on output queue 1;We enabled the expedite queue (priority-queue out). QuestionWith these settings we aim at forcing our synchronization packtes to precede other kind of traffic and go from Equipment 1 to Equipment 2 with minimum jitter.Unfortunately we experienced high jitter when both synchronization packets and other traffic are sent through the systems.
Trying to set-up a priority queue for Voice and Video traffic, below is the current ASA config. The WAN link is 6mb, trying to limit the Internet traffic to 4mb and save 2mb for the PQ, config belowTraffic just isn't hitting the PQ
priority-queue outside queue-limit 512 tx-ring-limit 200 ! class-map Video description Video match dscp af31
We have a failover pair of load balancers (non-Cisco) which are connected to each other via Catalyst 6509Es with Sup720 supervisor cards. Failover is achieved by the newly active load balancer GARPing all its service IP addresses with the relevant MAC address in order to update nearby ARP tables (failover GARPs are fired out by the load balancers at a rate of 200 per second). Failing over services between these load balancers has been found to be problematic, with numerous services not failing over in a timely manner.
Some of the load balanced networks involved are routed on the Sup720s, the rest are routed on FWSM modules in the same chassis. Problems occur only with VLANs routed on the Sup720s; all VLANs routed on the FWSMs fail over without issue.
Investigation has shown that this is due to a proportion of the ARP table entries not being updated in the Sup720 and (with the default 4 hour ARP table timeout) subsequently requiring a manual flush of the "wrong" ARP entries.
Testing by throwing GARPs at both the FWSM and the Sup720 has revealed the following:
we can quite happily throw ~200 GARPs per second at the FWSM and all the relevant ARP table entries are updated with the correct MAC address. This fits with the sucessful failovers for any FWSM routed networks. however, repeating the same test with a VLAN routed on the Sup720 results in GARPs being dropped by the Sup720. The Sup720 keeps discarding a proportion of GARPs until we drop the rate of GARP to <~75 per second. We're not rate-limiting ARP anywhere in the Sup720 - hardware or otherwise - and the FWSM handles the rate of GARP without issue. Is there any built-in restriction on the Sup720 we're not aware of that would cause the observed behaviour, is it configurable.
I have one specific trouble with Catalyst 3560CG switch and Stardom reserved industrial controller. Controller has 2 processors, one of them after(!) negotiation become main with IP 192.168.1.1, other one stays in reserve with IP 192.168.1.129. If main one shuts down, spare one becomes main with IP 192.168.1.1. This perfectly works with DLINK and MOXA level 2 switches, but fails with Catalyst 3560, because spare one cannot even connect to switch port. I don't know the exact reason of it, but i suspect, that it happens because of before negotiation both controllers have ip 192.168.1.1 (i know, that it is wrong and weird, but so it goes). First of them correctly connects to switch port, which goes up and passes pings. Second processor tries to connect to switch port, which goes several times up and very soon down, then processor lefts his tries to connect to anybody and become idle. The switch seems to some way check IPs of 1-st and 2-nd port packets, and shuts down 2-nd port. I have connected via hyperterminal and tried to disable some level 3 functions and checks of the switch, but this wasn't useful. how exactly(or at least approximately) should i configure a switch to make this bundle work?
I am having an issue pinpointing why my 2821 router is discarding so many packets when transferring data to our second site. The traffic flows from the local lan, to the router, where it is redirected via WCCP to a WAN optimization device, back to the router and over a GRE tunnel to the second site where the same process happens. The traffic does get there, but the LAN/Repeater router interfaces have around 20,000-60,000 input drops an hour. From the output below, it looks like traffic is being dropped by the RP.
I just restarted the router as a last resort, and here is what has accumulated in the last 30 min: FastEthernet0/0/1 is up, line protocol is up Hardware is Fast Ethernet, address is 0025.840c.7680 (bia 0025.840c.7680) [code]....
And CPU never goes above 40% 100 90 80 70 [code]...
I have been experiencing discards and no buffers errors on my 2950 switches. These switches are connected to two core switches 4948 and 3560G. I have noticed previous posts regrading these problems but have been unable to find any cases that were resolved. We are using 12.1(22)EA11 software on our 2950 switches.
Cisco 4500-X do not support egress queing on VLAN interfaces (SVI) which means cannot do a traffic-shapping, is there a work around via policing? I can police the traffic and then on the trunk interfaces do "per-port-per-VLAN" QoS but again only the policing not shapping so I was wondering what is the effect of "exceed-action transmit" command
policy-map SHAPE class class-default shape-average 8000000 Versus... policy-map POLICE class class-default police 8000000 4000 conform-action transmit exceed-action transmit
We're having some issues with a 3560 CPE. It's uplink is a GE fiber link, customer port is FE RJ45. We see a lot of TX-frames being dropped at the FE port,but none at the GE port. Even when the customer is only at ~50Mbit/s downstream traffic. When customer sending ~50Mbit/s upstream, there are no TX- drops at the GE link. Is this a normal behaviour? From what I know the physical medium shouldn't have any inpact on this since drops occurs in the port- asic, and not in physical transmission.Do the buffer-sizes between GE and FE differ? What could we do to optimize the flow and reduce drops? QoS is set to off and no modifications on the queues have been made on the interfaces.
We are upgrading from 3550 - 3560 switches.On the 3550's we have this on each interface: [code] The 3550's wont accept the wrr-queue commands. How to set these on the 3560's.
I have a 2921 where I am shaping some traffic based on sub net on my lan. I have applied the shaping policy to the lan interface in the outgoing direction.
Topology is as follows: ISP - ASA - ROUTER - LAN Policy map: Policy Map shape-lan [code]....
I am seeing a lot of no-buffer drops on the policy and I am wondering what the best solution is to solve this: Class-map: tc-class (match-any) 8730680 packets, 10803689863 bytes 5 minute offered rate 4453000 bps, drop rate 0 bps [code]....
Should I just be increasing the queue-limit or should I be changing something else?
One one of our Cisco 6509s I've globally enabled QoS and set a port to "trust cos". However when I look at the queueing for that interface, I notice that the receive queue thresholds have not changed to the default.
I'm kind of new to QoS so I'm not sure if I'm missing something.
We are using a Cisco 6509 running 12.2(33)SXI3 using the WS-X6724-SFP card. I thought I'd be seeing the default tail-drop thresholds, but instead I still see the defaults as if QoS were not enabled:
Queueing Mode In Rx direction: mode-cos Receive queues [type = 1q8t]: Queue Id Scheduling Num of thresholds
I am using 10Gig link to connect distribution and access switches. Since we are terminating the cables at multiple locations, we are also doing the link budgeting. The CISCO SFP 10G-LR module data sheet shows two values for the transmit power one of which is maximum power .5dBm and another is minimum power -8.2 dBm. Which value should I choose for the link budgeting.
I can see drops on the 6509 Queue for interface gi1/6 , qos is disabled globaly with qos disabled all packets are in one Queue using best effort my question is if I can see drops using the sh queueing int Gi1/6 command why I am not seeing any drops when I run the Sh int (interface number ) command. [code]
We tested a QoS in a Cisco 3750E, IOS: 12.2(58)SE2.Voice traffice in the correct Q without any problem, but all the others traffic the Defualt Q (0), tried to capture the traffic and tcp/udp port are correct.Any thing wrong with my ACL or DSCP - CoS?? ( that ACL works fin on 4500 and 6500) [code]
I have a Cisco 2960G switch and one of the ports was configured with srr-queue bandwidth limit 90 - I need to remove this bandwidth limiting from this interface. [code]
I can see drops on one of our busy L3 vlan in the input queue and are going up very frequently.System image file is "sup-bootflash:s72033-psv-mz.122-18.SXD3.binHardware = 6509
Recently we will add 2 new core Switch 3750X, these 2 equipments will manage the spanning tree ( root)my idea is to change the priority in order to make 1 or the 2 of them the root, my question is if i setup the same priority for both when one go down the other will assume the role of root in the spanning tree topology ?