You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Description @lguohan , We would eventually need a neighbor mgr. Atleast I could think of the following cases that may need some user-space handling. We could prioritize based on use-case.
Issue with gc threshold and flush of ARP entries
Periodic refresh of link-local entries.
Handle oper-down issue, where kernel does not delete the ARP entries but it might not be required to occupy HW resource for ARPs that belongs to a down interface
Handle admin-down scenario. If there are large number of ARP entries learnt on an interface, currently we rely on the the netlink notifications from kernel and this could be considerable depending on the interace. Instead a neigbhor manager could just act on the admin-down
Handle ECMP neighbor refresh. If one or two ECMP neigbhor members are resolved and programmed to ASIC, I don't think there is a mechanism to resolve the remaining members of ECMP. This could be handled by Nbr-mgr.
[Open]
6. FDB age-out in kernel and if ARP is not refreshed, kernel floods the IP packet to all member ports of Vlan.
7. Mgmt-VRF ARP reservation in case data-plane ARP fillsup the kernel threshold.
Steps to reproduce the issue:
1.
2.
3.
Describe the results you received:
Describe the results you expected:
Additional information you deem important (e.g. issue happens only occasionally):
**Output of `show version`:**
```
(paste your output here)
```
**Attach debug file `sudo generate_dump`:**
```
(paste your output here)
```
The text was updated successfully, but these errors were encountered:
Description
@lguohan , We would eventually need a neighbor mgr. Atleast I could think of the following cases that may need some user-space handling. We could prioritize based on use-case.
Issue with gc threshold and flush of ARP entries
Periodic refresh of link-local entries.
Handle oper-down issue, where kernel does not delete the ARP entries but it might not be required to occupy HW resource for ARPs that belongs to a down interface
Handle admin-down scenario. If there are large number of ARP entries learnt on an interface, currently we rely on the the netlink notifications from kernel and this could be considerable depending on the interace. Instead a neigbhor manager could just act on the admin-down
Handle ECMP neighbor refresh. If one or two ECMP neigbhor members are resolved and programmed to ASIC, I don't think there is a mechanism to resolve the remaining members of ECMP. This could be handled by Nbr-mgr.
[Open]
6. FDB age-out in kernel and if ARP is not refreshed, kernel floods the IP packet to all member ports of Vlan.
7. Mgmt-VRF ARP reservation in case data-plane ARP fillsup the kernel threshold.
Steps to reproduce the issue:
1.
2.
3.
Describe the results you received:
Describe the results you expected:
Additional information you deem important (e.g. issue happens only occasionally):
The text was updated successfully, but these errors were encountered: