01-13-2011 07:58 AM - edited 03-06-2019 02:58 PM
i have been experiencing a large amount of output queue drops on the 2960 and 3560 switch ports.
here are some of the things i have tried and the conclusions i have reached.
srr-queue bandwidth share 1 75 1 25 (original setting – the 75% is for queue 1 call signaling and 25% for queue 3 best effort )
srr-queue bandwidth share 1 50 1 50
srr-queue bandwidth share 1 25 1 75
after changing the queue setting I still experienced drops. In most cases it was less drops as the % increased for queue 3. in some cases it actually stayed the same or increased.
After removing qos I saw no drops on the 3560 switches. But on the 2960 switch I still saw a small amount of drops on a few ports. With qos all 11 10/100 ports had drops as high as 1000 over the same time period. With qos removed only 2 of the ports experienced drops and they were a small amount 1 and 91 drops.
Once I enabled qos on the switches the drops started occurring again on all 3 switches at a rapid rate.
i have read numerous posts on this issue. so i know i am not the only one seeing it. i contacted cisco tac about the issue and they continue to say that bursty traffic is causing all the drops. but what they can't account for is when removing qos almost all the drop go away.
has anyone else experienced this and if so were you able to fine a resolution. thank you.
i have also attached a show tech support file.
Solved! Go to Solution.
01-19-2011 05:35 PM
Hi Bob,
You are welcome! I am glad that I can help.
Regards,
Lei Tian
01-14-2011 12:57 PM
Upgrade your ios to 12.2(55) and let me know if that helps. It's helped others.
If it does help let me know. I have some switches with the same issue and that is what i've suggested to my boss.
https://cisco-support2.uat3.hosted.jivesoftware.com/thread/154488
01-14-2011 01:03 PM
hey danny,
according to the other posts i have read on this issue, the upgrade to 12.2(55) doesn't resolve the issue. according to one post i read a fix is scheduled for march 2011 in 12.2(58), according to the post. but i implemented the following qos changes recommended on this post and i am no longer seeing drops.
global changes
mls qos queue-set output 1 threshold 1 150 300 100 400
mls qos queue-set output 1 threshold 2 150 300 50 400
mls qos queue-set output 1 threshold 3 150 300 50 400
mls qos queue-set output 1 threshold 4 200 300 100 400
mls qos queue-set output 1 buffers 10 20 10 60
interface changes
srr-queue bandwidth share 10 20 10 60
hope that helps.
01-14-2011 01:16 PM
Unfortunately i'm seeing drops on my supervisor.
Supervisor TxQueue Drop Statistics
Queue 0: 37076
Queue 1: 0
Queue 2: 0
Queue 3: 1071053
Queue 4: 0
Queue 5: 0
Queue 6: 0
Queue 7: 0
Queue 8: 769441
Queue 9: 0
Queue 10: 1027404
Queue 11: 0
Queue 12: 0
Queue 13: 0
Queue 14: 38879547
Queue 15: 0
01-14-2011 02:23 PM
Drops are most likely due to applications trying to transmit faster than interface speed, thing than in networking has been proven impossible so far.
01-14-2011 02:31 PM
The main queue that i'm having issues is the dstats queue, which is the drop stats queue. That's all i know about it.
I'm hoping on the ios upgrade.
cpu-queue-frames retrieved dropped invalid hol-block stray
----------------- ---------- ---------- ---------- ---------- ----------
rpc 105040525 0 0 0 0
stp 9346673 0 9343664 0 0
ipc 11447788 0 0 0 0
routing protocol 26482989 0 0 0 0
L2 protocol 547553 0 43 0 0
remote console 0 0 0 0 0
sw forwarding 0 0 0 0 0
host 2231757 0 0 0 0
broadcast 10729955 0 0 0 0
cbt-to-spt 0 0 0 0 0
igmp snooping 19845858 0 0 0 0
icmp 0 0 0 0 0
logging 0 0 0 0 0
rpf-fail 0 0 0 0 0
dstats 78486304 0 0 0 0
cpu heartbeat 46374282 0 0 0 0
Supervisor TxQueue Drop Statistics
Queue 0: 37160
Queue 1: 0
Queue 2: 0
Queue 3: 1071053
Queue 4: 0
Queue 5: 0
Queue 6: 0
Queue 7: 0
Queue 8: 769528
Queue 9: 0
Queue 10: 1027404
Queue 11: 0
Queue 12: 0
Queue 13: 0
Queue 14: 38908191
Queue 15: 0
01-15-2011 11:06 AM
Hi
I have the same problem of packet dropped on a Cisco 3560E (10/100/1000 ports and uplink at 10GE) with IOS 12.2.(53)S2
In my scenario the 3560 has a portchannel with two GE ports connected to a server and this server does the backup of all documents,e-mail ecc ecc of the customers. I have no qos enabled so all the switch works with default value.
I notice the dropped output an interface only (the traffic was very high near 800Mb/s) while the other interace had no traffic in output so I changed the load-balance method for source-ip instead of source-mac
The situation seems to work fine because the switch started to use both the interfece but , after a while , the packet dropping started again even if the traffic was at 300/400Mb/s
It seems that I have a problem with bursty traffic
Must I enable the mls qos and srr to solve this problem?
The other server connect to the same switch haven't any packet dropped
If I configure the mls qos could I have problem with these servers?
thanks
Riccardo
01-15-2011 01:55 PM
As mentioned above. Drops are due to burst of traffic, are normal and there is nothing you can do about it.
01-18-2011 04:38 AM
i have been working w/ cisco tac on this issue and their response has been that bursty traffic is causing the issue.
my response to that is below
srr-queue bandwidth share 1 75 1 25 (original setting – the 75% is for queue 1 call signaling and 25% for queue 3 best effort )
srr-queue bandwidth share 1 50 1 50
srr-queue bandwidth share 1 25 1 75
after changing the queue setting I still experienced drops. In most cases it was less drops as the % increased for queue 3. in some cases it actually stayed the same or increased.
After removing qos I saw no drops on the 3560 switches in hw. But on the 2960 switch in admin I still saw a small amount of drops on a few ports. With qos all 11 10/100 ports had drops. With qos removed only 2 of the ports experienced drops and they were a small amount 1 and 91 drops.
Once I enabled qos on the switches the drops started occurring again on all 3 switches at a rapid rate.
after tuning the qos with commands i mentioned earlier and letting it run for a few days, i am seeing no drops on the 3560 switches that support only 10/100. i am still seeing a few drops on the 2960s switches that support 10/100/1000. only ports running at 10/100 are getting drops and now it's only a few like 20 or 30 as opposed to thousands of drops int the same time period.
i am going to test the new qos commands on some other 2960 switches i have that support only 10/100 and users connecting to these switches have complained about slowness and see if that helps. i will postr the results in a few days.
01-19-2011 12:43 PM
this is cisco tac's response to this issue
Hi Bob,
Thanks for the update!
I would like to tell you that the drops are due to traffic bursts on the interfaces.
Tweaking of threshold on the queues is a good option as it will not drop the packets early and would rather allot more buffers to the queues.
If this also does not work than you can wait till March 2011 for the new IOS release and upgrade the switch to that IOS.
Thank you for your time and understanding.
Since I am familiar with your service request, I would like to retain ownership to avoid any delays in resolving your issue.
Thanks & Regards
Sumit Sharma
Customer Support Engineer - LAN Switching
Cisco Systems, Inc.
since i have tuned the qos settings with recommendations provided by lei on this post, i have significantly reduced the amount of drops. in some cases none at all. please see the attached file for a 24 hour time frame sample. lei, the last recommendation for increasing the max threshold for queue 4 to 1000 worked like a charm. thank you.
at this point the qos recommendations have resolved most of my drop issues. i also spoke to users on these switches and they are also seeing a performance increase. i am still seeing some drops but i believe most of these are related to bursty traffic and possible issues with the ios (high cpu utilization). when the new ios is released i am definitely going to take a look at upgrading.
thank again for everyone's help and insight.
05-11-2011 08:06 AM
hey danny,
according to the other posts i have read on this issue, the upgrade to 12.2(55) doesn't resolve the issue. according to one post i read a fix is scheduled for march 2011 in 12.2(58), according to the post. but i implemented the following qos changes recommended on this post and i am no longer seeing drops.
global changes
mls qos queue-set output 1 threshold 1 150 300 100 400
..................................................................................
Hi
does anyone tried the 12.2(58)SE1 to solve the problem of this post?
Please let me know a feedback..
Regards Riccardo
05-11-2011 03:23 PM
I must be the first person to use the new 12.2(58)SE1. So far so good. Testing is good. I'm now rolling it into the "wild".
05-23-2011 05:21 PM
Hi there, I have some 2960 switches running the LAN lite image that are also experiancing drops on queue 4 for best effort traffic when QoS is enabled.
We have upgraded the IOS to 12.2(58)SE1 but it hasn't resolved the issue.
sh mls qos int fa0/1 buffers
FastEthernet0/1
The port is mapped to qset : 1
The allocations between the queues are : 25 25 25 25
sh platform port-asic stats drop fa0/1
Interface Fa0/1 TxQueue Drop Statistics
Queue 0
Weight 0 Frames 0
Weight 1 Frames 0
Weight 2 Frames 0
Queue 1
Weight 0 Frames 0
Weight 1 Frames 0
Weight 2 Frames 0
Queue 2
Weight 0 Frames 0
Weight 1 Frames 0
Weight 2 Frames 0
Queue 3
Weight 0 Frames 0
Weight 1 Frames 0
Weight 2 Frames 74007
the show mls qos queue-set command does not exist.
Unfortunately the LAN Lite image doesn't look to provide the following commands in order to manually modify the buffers etc.
mls qos queue-set output 1 threshold
mls qos queue-set output 1 buffers
srr-queue bandwidth share
Are there any other options avilable in a LAN Lite image ?
05-23-2011 05:30 PM
LAN Lite won't do QoS.
05-23-2011 06:18 PM
As far as I'm aware basic QoS is provided in terms of egress queueing and priority. It doesn't have auto QoS.
I need to amend the buffers defined for egress as per previous recomendations.
This is a PoE switch so is expected to carry VoIP traffic, which i wouldn't expect to be at the expense of the majority of the traffic.
Can anyone shed some light (excuse the pun).
sh mls qos int fa0/1
FastEthernet0/1
trust state: trust cos
trust mode: trust cos
trust enabled flag: ena
COS override: dis
default COS: 0
sh mls qos int fa0/1 statistics
FastEthernet0/1 (All statistics are in packets)
dscp: incoming
-------------------------------
0 - 4 : 113392 0 0 0 0
5 - 9 : 0 0 0 0 0
10 - 14 : 0 0 0 0 0
15 - 19 : 0 0 0 0 0
20 - 24 : 0 0 0 0 1418
25 - 29 : 0 0 0 0 0
30 - 34 : 0 0 0 0 0
35 - 39 : 0 0 0 0 0
40 - 44 : 0 0 0 0 0
45 - 49 : 0 16866 0 0 0
50 - 54 : 0 0 0 0 0
55 - 59 : 0 0 0 0 0
60 - 64 : 0 0 0 0
dscp: outgoing
-------------------------------
0 - 4 : 160378 0 0 0 0
5 - 9 : 0 0 0 0 0
10 - 14 : 0 0 0 0 0
15 - 19 : 0 0 0 0 0
20 - 24 : 0 0 0 0 1105
25 - 29 : 0 74 0 0 0
30 - 34 : 0 0 0 0 0
35 - 39 : 0 0 0 0 0
40 - 44 : 0 0 0 0 0
45 - 49 : 0 16942 0 2124 0
50 - 54 : 0 0 0 0 0
55 - 59 : 0 0 0 0 0
60 - 64 : 0 0 0 0
cos: incoming
-------------------------------
0 - 4 : 113901 0 0 1418 0
5 - 7 : 16866 0 0
cos: outgoing
-------------------------------
0 - 4 : 228734 0 0 0 0
5 - 7 : 0 575 7
output queues enqueued:
queue: threshold1 threshold2 threshold3
-----------------------------------------------
queue 0: 2 0 0
queue 1: 0 10433 12542
queue 2: 0 0 0
queue 3: 0 0 218801
output queues dropped:
queue: threshold1 threshold2 threshold3
-----------------------------------------------
queue 0: 0 0 0
queue 1: 0 0 0
queue 2: 0 0 0
queue 3: 0 0 2657
sh ver
Cisco IOS Software, C2960 Software (C2960-LANLITEK9-M), Version 12.2(58)SE1, RELEASE SOFTWARE (fc1)
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide