This post is more than 5 years old
9 Posts
0
116181
Uplink port on PowerConnect 8024F went to state D-Down
Hello everyone,
we had a strange behaviour on one of our PowerConnect 8024F switches. The uplink port of the device (connected to a 10Gig port on a Cisco Catalyst 6513) went down. The output of 'show interfaces status' looked like that:
Port Description Duplex Speed Neg Link Flow Control
State Status
--------- ------------------------- ------ ------- ---- ------ ------------
Te1/0/1 cat6500-rz-1 (te11/2) Full 10000 Off D-Down Inactive
Te1/0/2 pc8024-ub-1 (te1/0/1) Full 10000 Off Up Active
Te1/0/3 pc8024-uv-1 (te1/0/1) Full 10000 Off Up Active
Te1/0/4 pc5500-naf02n-1 (te1/0/1) Full 10000 Off Up Active
Te1/0/5 pc5500-naf04n-1 (te1/0/1) Full 10000 Off Up Active
Te1/0/6 pc5500-naf03-1 (te1/0/2) Full 10000 Off Up Active
Te1/0/7 pc8024-mensa-1 (te1/0/1) Full 10000 Off Up Active
Te1/0/8 defekt N/A Unknown Auto Down Inactive
Te1/0/9 pc5500-naf03-2 (te1/0/1) Full 10000 Off Up Active
Te1/0/10 pc5500-naf03-3 (te1/0/1) Full 10000 Off Up Active
...
...and so on. Firmware is 5.1.2.3. We rebooted the device and it went back to normal behaviour.
Now, can anyone explain to me what the state "D-Down" means? I did not find anything in the documentation nor did using google have any success.
The port configuration looks like the following:
description "cat6500-rz-1 (te11/2)"
mtu 9216
switchport mode general
switchport general allowed vlan add 4-5,14,22-27,29-33,36,39-41,43 tagged
switchport general allowed vlan add 47,51-52,54-56,58-60,63-64,66,75,78-79 tagged
switchport general allowed vlan add 81,86-90,93-99,107,120,122 tagged
switchport general allowed vlan add 124,132-135,137,140,142-143,147-152,154 tagged
switchport general allowed vlan add 156-157,160,165-166,171,173-175 tagged
switchport general allowed vlan add 177-178,180-181,183,189,192,195-196 tagged
switchport general allowed vlan add 199,201,203-205,207-208,210-212 tagged
switchport general allowed vlan add 219-222,224-225,227-233,235-244,246 tagged
switchport general allowed vlan add 248-249,253,256,258-260,262-267 tagged
switchport general allowed vlan add 269-272,274-276,278-279,282,284,286-288 tagged
switchport general allowed vlan add 292,295-298,323,344,347,349 tagged
switchport general allowed vlan add 354,362,366,370,372,378-380,382-383,385-386 tagged
switchport general allowed vlan add 391,399,405,431,443,447 tagged
switchport general allowed vlan add 475-476,485,488,511,522,584,700-703,710-714 tagged
switchport general allowed vlan add 725,734,838,854-855,909 tagged
switchport general allowed vlan add 950,961-962,981,999,1234,1236,2205,2218-2220 tagged
switchport general allowed vlan add 2222,2233,2414,2417 tagged
switchport general allowed vlan add 2506-2507,2519,2521,2621,2699 tagged
mac access-group PVST-Filter out 1
lldp transmit-tlv sys-name sys-cap
lldp transmit-mgmt
The used mac access-group PVST-Filter is:
mac access-list extended PVST-Filter
deny any 0100.0CCC.CCCD 0000.0000.0000
permit any any
exit
Thank you for reading and any suggestions!
Robin
DELL-Josh Cr
Moderator
Moderator
•
8.5K Posts
0
April 22nd, 2015 13:00
It is all spanning-tree related. D-down stands for diagnostically down, which occurs if more than 15 bpdus per second are received for 3 seconds. The switch shuts the port down. USL worker task is traffic related so a big spanning tree change would cause that to be really high.
DELL-Josh Cr
Moderator
Moderator
•
8.5K Posts
0
April 22nd, 2015 11:00
Hi Robin,
Is the port in a port channel ? Does it show a more detailed error in the logs? Show logging file
Page 1677 http://downloads.dell.com/Manuals/Common/PowerConnect_8024-and-F-CLIRef_en-us.pdf
schrorg
9 Posts
0
April 22nd, 2015 13:00
Hi Josh,
no, it is not a port-channel.
The log file shows only the following:
Persistent Logging : enabled
Persistent Log Count : 6278
<184> APR 22 20:56:22 10.11.0.123-1 MGMT_ACAL[81238496]: macal.c(742) 19049938 %% Number of packets denied by management ACLs = 94662
<184> APR 22 20:41:20 10.11.0.123-1 MGMT_ACAL[81238496]: macal.c(742) 19048152 %% Number of packets denied by management ACLs = 94647
<184> APR 22 20:26:20 10.11.0.123-1 MGMT_ACAL[81238496]: macal.c(742) 19046485 %% Number of packets denied by management ACLs = 94632
<184> APR 22 20:11:19 10.11.0.123-1 MGMT_ACAL[81238496]: macal.c(742) 19044877 %% Number of packets denied by management ACLs = 94617
...
Whereas the output of 'show logging' shows also a lot of the usual STP topology changes, i.e.:
<189> APR 22 18:50:56 10.11.0.86-1 TRAPMGR[217599600]: traputil.c(638) 326358 %% Spanning Tree Topology Change Received: MSTID: 7 Te1/0/13
<189> APR 22 18:50:56 10.11.0.86-1 TRAPMGR[217599600]: traputil.c(638) 326357 %% Spanning Tree Topology Change Received: MSTID: 6 Te1/0/13
<189> APR 22 18:50:56 10.11.0.86-1 TRAPMGR[217599600]: traputil.c(638) 326356 %% Spanning Tree Topology Change Received: MSTID: 5 Te1/0/13
<189> APR 22 18:50:56 10.11.0.86-1 TRAPMGR[217599600]: traputil.c(638) 326355 %% Spanning Tree Topology Change Received: MSTID: 4 Te1/0/13
<189> APR 22 18:50:56 10.11.0.86-1 TRAPMGR[217599600]: traputil.c(638) 326354 %% Spanning Tree Topology Change Received: MSTID: 3 Te1/0/13
<189> APR 22 18:50:56 10.11.0.86-1 TRAPMGR[217599600]: traputil.c(638) 326353 %% Spanning Tree Topology Change Received: MSTID: 2 Te1/0/13
<189> APR 22 18:50:56 10.11.0.86-1 TRAPMGR[217599600]: traputil.c(638) 326352 %% Spanning Tree Topology Change Received: MSTID: 1 Te1/0/13
<189> APR 22 18:50:56 10.11.0.86-1 TRAPMGR[217599600]: traputil.c(638) 326351 %% Spanning Tree Topology Change Received: MSTID: 0 Te1/0/13
Indeed, we narrowed down the problem to the spanning-tree today. About 2000 switches in one MST region seem to be too much (in contrast to what Cisco means in their documentation...), so we reverted a change from about three weeks ago and disabled spanning-tree on some uplink ports again.
Before the uplink port went into the state "D-Down", the CPU load on the switch rised to arount 90% and everything stuttered. It was the "USL Worker Task" using the CPU. We did not find any information on what this task does, but now that we have much less topology changes in the region, the process entirely vanished from the output of 'show process cpu' and everything went back to normal.
Is there any information on what "D-Down" really means and what the "USL Worker Task" does?
Robin
schrorg
9 Posts
0
April 23rd, 2015 07:00
Hi Josh,
thank you for your assistance. Everything is back to normal now.
It would have been helpful to find anything about "D-Down" in the documentation. :-/
Is it possible to configure the values at which ports go to D-Down state?
Robin
DELL-Josh Cr
Moderator
Moderator
•
8.5K Posts
0
April 23rd, 2015 10:00
There is not a way to change the timeout, but you can allow bpdu flooding. Page 744 http://downloads.dell.com/Manuals/Common/PowerConnect_8024-and-F-CLIRef_en-us.pdf
Spanning-tree bpdu flooding
schrorg
9 Posts
0
June 13th, 2016 03:00
The port state does not recover automatically, you have to "shut" and "no shut" the port.
ponyma2015
6 Posts
0
June 13th, 2016 03:00
Dear Sir
After loop condition was removed , it seems that the link status still "D-DOWN"
how can I do to let it recovery to normal state automatically
thanks...