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
fix: [WIN-NPM] process updatePods in fifo order (#1856)
* process updatePods in fifo order
* fix lint
* better UT
* comments and better naming
* stop skipping UTs
* fix lint
* redesign
* dequeue returns nil when cache is empty
* Revert "dequeue returns nil when cache is empty"
This reverts commit 3e8d187.
* requeue if node name has changed
* Revert "Revert "dequeue returns nil when cache is empty""
This reverts commit 3f5f99d.
* UT for nil result from dequeue
// should never happen because of isEmpty check above and lock on updatePodCache
224
+
metrics.SendErrorLogAndMetric(util.DaemonDataplaneID, "[DataPlane] failed to dequeue pod while applying the dataplane")
225
+
// break to avoid infinite loop (something weird happened since isEmpty returned false above)
226
+
break
227
+
}
228
+
229
+
iferr:=dp.updatePod(pod); err!=nil {
244
230
// move on to the next and later return as success since this can be retried irrespective of other operations
245
-
metrics.SendErrorLogAndMetric(util.DaemonDataplaneID, "failed to update pod while applying the dataplane. key: [%s], err: [%s]", podKey, err.Error())
231
+
metrics.SendErrorLogAndMetric(util.DaemonDataplaneID, "failed to update pod while applying the dataplane. key: [%s], err: [%s]", pod.PodKey, err.Error())
0 commit comments