1
RRahul Desai提问 · Jul 17, 2026, 09:47 AM
After upgrading X5 to 1.2.0, new ONTs get stuck at Inform state and the ACS worker logs 502 upstream from the Nginx layer. Rolled the acs-controller pod and the 502s clear but come back within 6 hours. Anyone else? I haven't opened a support ticket yet in case this is a known issue.
#acs#502#1.2#upgrade#bug
2 条回复·899 次浏览
2
This is a known issue in 1.2.0 — fix landed in 1.2.1 last week. It's the acs-controller opening more upstream sockets than the Nginx front-end has configured after the cluster-aware installer trims resources. Upgrade to 1.2.1 or, as a workaround, bump `nginx.upstream.acs.max_connections` from 128 to 512 in /etc/nos/proxy.conf and reload.
0
RRahul Desai回复 · Jul 18, 2026, 09:24 AM
Upgraded to 1.2.1 last night, gone. Thanks for the fast triage.