Output Drops in Cisco 2960

[Sgj]

Hello,

I found Output Drops in some Interfaces in Cisco 2960 series and want to solve this.

But when trying to type the command 'qos queue-softmax-multiplier 1200' , it doesn't work.

I already try this command in other series, it works.

Does anybody know the correct command of 'qos queue-softmax-multiplier 1200' in the 2960 series?

Thanks in advanced

[Georg Pauwen]

Hello,

post the running configuration (sh run) of the switch, as I am not sure what you have configured currently...

[Sgj]

Hello,

Here I attach the file

[Georg Pauwen]

Hello,

configure the below values and monitor again:

no mls qos queue-set output 2 threshold 1 149 149 100 149
no mls qos queue-set output 2 threshold 2 118 118 100 235
no mls qos queue-set output 2 threshold 3 3100 3100 100 3200
no mls qos queue-set output 2 threshold 4 42 72 100 242

mls qos queue-set output 2 threshold 1 3200 3200 100 3200
mls qos queue-set output 2 threshold 2 3200 3200 100 3200
mls qos queue-set output 2 threshold 3 3200 3200 100 3200
mls qos queue-set output 2 threshold 4 3200 3200 100 3200

[Sgj]

Hello,

After input the command and monitor again, there's still drop the packet

AS02-3#sh int fa0/14
FastEthernet0/14 is up, line protocol is up (connected)
Hardware is Fast Ethernet, address is dc7b.94df.440e (bia dc7b.94df.440e)
MTU 1500 bytes, BW 100000 Kbit/sec, DLY 100 usec,
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA, loopback not set
Keepalive set (10 sec)
Full-duplex, 100Mb/s, media type is 10/100BaseTX
input flow-control is off, output flow-control is unsupported
ARP type: ARPA, ARP Timeout 04:00:00
Last input 00:00:29, output 00:00:01, output hang never
Last clearing of "show interface" counters 16:53:07
Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 31
Queueing strategy: fifo
Output queue: 0/40 (size/max)
5 minute input rate 5000 bits/sec, 1 packets/sec
5 minute output rate 21000 bits/sec, 10 packets/sec
260056 packets input, 45189884 bytes, 0 no buffer
Received 31702 broadcasts (25499 multicasts)
0 runts, 0 giants, 0 throttles
0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
0 watchdog, 25499 multicast, 0 pause input
0 input packets with dribble condition detected
768466 packets output, 477368797 bytes, 0 underruns
0 output errors, 0 collisions, 0 interface resets
0 unknown protocol drops
0 babbles, 0 late collision, 0 deferred
0 lost carrier, 0 no carrier, 0 pause output
0 output buffer failures, 0 output buffers swapped out
AS02-3#sh int fa0/24
FastEthernet0/24 is up, line protocol is up (connected)
Hardware is Fast Ethernet, address is dc7b.94df.4418 (bia dc7b.94df.4418)
MTU 1500 bytes, BW 100000 Kbit/sec, DLY 100 usec,
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA, loopback not set
Keepalive set (10 sec)
Full-duplex, 100Mb/s, media type is 10/100BaseTX
input flow-control is off, output flow-control is unsupported
ARP type: ARPA, ARP Timeout 04:00:00
Last input 00:00:21, output 00:00:00, output hang never
Last clearing of "show interface" counters 16:53:08
Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 1433
Queueing strategy: fifo
Output queue: 0/40 (size/max)
5 minute input rate 34000 bits/sec, 9 packets/sec
5 minute output rate 134000 bits/sec, 27 packets/sec
439399 packets input, 101135045 bytes, 0 no buffer
Received 9418 broadcasts (5105 multicasts)
0 runts, 0 giants, 0 throttles
0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored
0 watchdog, 5105 multicast, 0 pause input
0 input packets with dribble condition detected
881284 packets output, 236303270 bytes, 0 underruns
0 output errors, 0 collisions, 0 interface resets
0 unknown protocol drops
0 babbles, 0 late collision, 0 deferred
0 lost carrier, 0 no carrier, 0 pause output
0 output buffer failures, 0 output buffers swapped out

[Joseph W. Doherty]

Lots of non-default values, I see.

What's the history of your current settings?

Reason I ask, before making suggestions on changes, it would help to understand the goals (assuming beyond just not dropping packets)?

BTW, what Georg suggests using queue-set 2 and 100 for the 3rd parameter, I've often had good results (on 3750) just using queue-set 1 with a "larger" common pool allocation (which is accomplished by reducing the 3rd parameter, to a minimum or near minimum value).  Doing so allows ports to obtain buffers that otherwise would be unused but "reserved" to other active interfaces.  Doing this also often avoids the need to reallocate buffer allocations per queue, too.

I.e.:

mls qos queue-set output 1 threshold 1 3200 3200 1 3200
mls qos queue-set output 1 threshold 2 3200 3200 1 3200
mls qos queue-set output 1 threshold 3 3200 3200 1 3200
mls qos queue-set output 1 threshold 4 3200 3200 1 3200

mls qos queue-set output 2 threshold 1 3200 3200 1 3200
mls qos queue-set output 2 threshold 2 3200 3200 1 3200
mls qos queue-set output 2 threshold 3 3200 3200 1 3200
mls qos queue-set output 2 threshold 4 3200 3200 1 3200

mls qos queue-set output 1 buffers 25 25 25 25
mls qos queue-set output 2 buffers 25 25 25 25

[Sgj]

Hi Joseph, thank you for the advice

After I input this,

mls qos queue-set output 1 threshold 1 3200 3200 1 3200
mls qos queue-set output 1 threshold 2 3200 3200 1 3200
mls qos queue-set output 1 threshold 3 3200 3200 1 3200
mls qos queue-set output 1 threshold 4 3200 3200 1 3200

mls qos queue-set output 2 threshold 1 3200 3200 1 3200
mls qos queue-set output 2 threshold 2 3200 3200 1 3200
mls qos queue-set output 2 threshold 3 3200 3200 1 3200
mls qos queue-set output 2 threshold 4 3200 3200 1 3200

mls qos queue-set output 1 buffers 25 25 25 25
mls qos queue-set output 2 buffers 25 25 25 25

There's not showing the mls qos queue-set output x buffers x x x x

[Joseph W. Doherty]

It (mls qos queue-set output # buffers 25 25 25 25) is no longer shown because that's the default.

How's the drop count doing with the changes I suggested?

BTW, it's not always possible to totally eliminate drops, more so on a 2960 vs. 3560/3750 (because the latter have more physical buffer RAM).

Unsure about the 2960s, but on the 3560/3750s they partition physical buffer RAM for each bank of 24 copper ports (if 48 port switch) and for the uplink ports (if provided).  Physical resource allocations to groups of ports is not unusual, so it can sometimes make a difference how you distribute connections, i.e. insuring either usage is equal across all the physical resource allocations or insure your "critical" ports don't share the same physical resource as "non-critical" ports.  (For the latter, you're attempting to insure your "critical" ports have no, or minimal, drops, but at the possibility, or increase, of drops for your "non-critical" ports.)

[Georg Pauwen]

On a side note, I wonder if the drops disappear when you just disable mls altogether (no mls qos)...

[Sgj]

Hello,

I once entered this command, but there's still mls qos and its qos list

[Sgj]

Hello,

There are still showing the Drop packet

[Joseph W. Doherty]

As I noted, it can be difficult to completely eliminate all drops.  However, have you seen any change in the drop percentage?

[Joseph W. Doherty]

What Georg is describing is what you likely need to consider for a 2960, which, more or less, uses the same QoS architecture as the 3560/3750 series.  (You might also search the forums on QoS and/or drops, and 3750 for more information that should also apply to your 2960 too.)

[paul driver]

Hello
On those two interfaces could you also try the following to negate the default srr shape weight that will police them down to 4mbs.

int x/x
no priority-queue out
srr-queue bandwidth shape 0 0 0 0

[Sgj]

Hello,

Maybe I'll try this

Thank you for the idea

[Sgj]

Hello,

It doesn't work too