On ACX710/ACX5448 platforms in the layer2 circuit scenario, if the primary path is down due to any reason, the backup circuit takes the mastership and ideally, once the primary path is up, mastership is to be reverted. But, in rare cases, the mastership gets reverted but the label might get lost from PFE and results in traffic loss for the l2circuit connection. However, deactivating and activating the affected l2circuit configuration could help in restoring the services.

This issue might be seen if the following conditions are met:
* On ACX710/ACX5448 platforms 
* Primary l2circuit is down due to any reason followed by immediate role change (for example, flapping)

The issue has happened due to some sync issue between RE and PFE after the trigger for mastership role change. The issue is expected only during sudden role changes, for example, during link flapping. Please find the related CLIs below.

user@device#set protocols l2circuit  -----> mandatory CLI to be present for the issue to happen
user@device>show l2circuit connections  -----> to view l2circuit sessions
user@device>show log messages  ----->sample output once the issue has hit the network
<date> <time> <device> fpc<slot no> dnx_rt_tag_ccc_entry_hw_remove:UNI VPWS port_add failed RT: 18 NH: 656 (-6:Table full)
<date> <time> <device> fpc<slot no>  dnx_rt_vswitch_cross_connect_add_del: dnx_rt_vswitch_cross_connect_add_del:cross-connect delete failed for IFL xx(UNI xxxxxxx,NNI xxxxxxxx) (-1:Internal error)
<date> <time> <device> fpc<slot no>  dnx_rt_ccc_eth_vpws_vpn_nni_port_add:NNI VPWS port_del failed  NH nh_ptr=xxxxxxx(xxxxxxx),uc_nh_ptr=xxxxxxxx(677) error - (-7:Entry not found)

Issue is fixed in the list of releases in PR1661802

Workaround:
deactivating and activating the affected l2circuit configuration could help in restoring the services.