
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
03-20-2023 10:45 PM - edited 03-20-2023 10:48 PM
Hi,
We have four Nexus 93180YC-EX running NXOS 7.0(3)I4(6) and an uptime of > 2100 days. I need to update them to the latest recommended NXOS version.
Management interface is a VLAN interface. The dedicated management interface is used for vPC peer keep-alive.
Copying a software image of circa 950 MB size takes about 2 hours, no matter whether I copy it via HTTP or SCP or SFTP from a host in the same datacenter. Another copy did not finish after 12 hours. We have a bunch of newer Nexus switches that came with NXOS 9.3.x and there same copy was much faster.
nexus9k# copy http://1.2.3.4/cisco/nexus/93180YC-EX/nxos.7.0.3.I7.10.bin bootflash:
Enter vrf (If no input, current vrf 'default' is considered):
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
16 940M 16 151M 0 0 136k 0 1:57:23 0:18:57 1:38:26 105k
Is it a software bug or due to the long uptime?
Regards,
Bernd
Solved! Go to Solution.
- Labels:
-
Other Switches
Accepted Solutions
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
03-21-2023 01:20 AM
- FYI : https://bst.cisco.com/bugsearch/bug/CSCvh45935
M.
-- Each morning when I wake up and look into the mirror I always say ' Why am I so brilliant ? '
When the mirror will then always repond to me with ' The only thing that exceeds your brilliance is your beauty! '
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
03-21-2023 01:20 AM
- FYI : https://bst.cisco.com/bugsearch/bug/CSCvh45935
M.
-- Each morning when I wake up and look into the mirror I always say ' Why am I so brilliant ? '
When the mirror will then always repond to me with ' The only thing that exceeds your brilliance is your beauty! '

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
03-21-2023 01:51 AM
Thanks. With NXOS NXOS 7.0(3)I4(6) the default copp policy is:
nexus9k# show run all | section copp-system-p-policy-strict
policy-map type control-plane copp-system-p-policy-strict
class copp-system-p-class-l3uc-data
set cos 1
police cir 800 kbps bc 32000 bytes conform transmit violate drop
class copp-system-p-class-critical
set cos 7
police cir 36000 kbps bc 1280000 bytes conform transmit violate drop
class copp-system-p-class-important
set cos 6
police cir 2500 kbps bc 1280000 bytes conform transmit violate drop
class copp-system-p-class-openflow
set cos 5
police cir 1000 kbps bc 32000 bytes conform transmit violate drop
class copp-system-p-class-multicast-router
set cos 6
police cir 2600 kbps bc 128000 bytes conform transmit violate drop
class copp-system-p-class-multicast-host
set cos 1
police cir 1000 kbps bc 128000 bytes conform transmit violate drop
class copp-system-p-class-l3mc-data
set cos 1
police cir 2400 kbps bc 32000 bytes conform transmit violate drop
class copp-system-p-class-normal
set cos 1
police cir 1400 kbps bc 32000 bytes conform transmit violate drop
class copp-system-p-class-ndp
set cos 6
police cir 1400 kbps bc 32000 bytes conform transmit violate drop
class copp-system-p-class-normal-dhcp
set cos 1
police cir 1300 kbps bc 32000 bytes conform transmit violate drop
class copp-system-p-class-normal-dhcp-relay-response
set cos 1
police cir 1500 kbps bc 64000 bytes conform transmit violate drop
class copp-system-p-class-normal-igmp
set cos 3
police cir 3000 kbps bc 64000 bytes conform transmit violate drop
class copp-system-p-class-redirect
set cos 1
police cir 280 kbps bc 32000 bytes conform transmit violate drop
class copp-system-p-class-exception
set cos 1
police cir 150 kbps bc 32000 bytes conform transmit violate drop
class copp-system-p-class-exception-diag
set cos 1
police cir 150 kbps bc 32000 bytes conform transmit violate drop
class copp-system-p-class-management
set cos 2
police cir 10000 kbps bc 32000 bytes conform transmit violate drop
class copp-system-p-class-monitoring
set cos 1
police cir 150 kbps bc 128000 bytes conform transmit violate drop
class copp-system-p-class-l2-unpoliced
set cos 7
police cir 50 mbps bc 8192000 bytes conform transmit violate drop
class copp-system-p-class-undesirable
set cos 0
police cir 200 kbps bc 32000 bytes conform transmit violate drop
class copp-system-p-class-fcoe
set cos 6
police cir 1060 kbps bc 128000 bytes conform transmit violate drop
class copp-system-p-class-nat-flow
set cos 7
police cir 800 kbps bc 64000 bytes conform transmit violate drop
class copp-system-p-class-l2-default
set cos 0
police cir 400 kbps bc 32000 bytes conform transmit violate drop
class class-default
set cos 0
police cir 400 kbps bc 32000 bytes conform transmit violate drop
service-policy input copp-system-p-policy-strict
The planned recommended upgrade path is NXOS 7.0(3)I4(6) → 7.0(3)I7(3) → 7.0(3)I7(10) → 9.3(11) → 10.2(4)M. From the bug description this is fixed in 7.0(3)I7(3).
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
06-14-2023 02:22 AM
I have similar issue, but policy values seems to be corrected, in compare with bug description. We have NX-OS 9.3(7).
class copp-system-p-class-management
set cos 2
police cir 36000 kbps bc 512000 bytes conform transmit violate drop
SCP transfer starts at 2MB/s, then after few seconds drop to ~60KB/s. Upload image take ages with this speed. Source (windows server) and destination (switch) are in same subnet.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
10-13-2023 01:20 PM
You can use the “use-kstack” syntax with the copy command on the nexus switches to increase the speed of the copy commands. See attached documentation for details.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
10-13-2023 03:30 PM
@dgaunt did you note in OP: ". . . I copy it via HTTP or SCP or SFTP from a host in the same datacenter."
and in your reference: "This option can be beneficial when copying files from remote servers that are multiple hops from the switch."
So, might not make much of a difference in OP's case.
Also, unclear whether the alternative network stack would bypass CoPP, or the noted bug, again, might not make a difference.
That said, should be easy enough to try, and I suspect, unlikely to do worst or have other adverse effect.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
05-24-2024 01:40 PM - edited 05-24-2024 01:42 PM
I'm having this problem too on four N9K-C9336C-FX2 running 10.2.6M via the mg0 port on the management vrf. The file transfers seem to be capped at ~320KB/s per switch. If I run two switches at once, they both hit around 320KB/s so its not any other limiter along the way that I can see.
I've tried http and sftp. It's not a local server but its connected via 1gbps w 1gpbs switching and into a 10gbps internet connection that no one else is using.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
05-24-2024 02:04 PM
If the actual file transfer is via the management port, the port might be the issue. Sometimes such ports are not hardware supported like the data ports.
