rootnomad commented on code in PR #14038:
URL: https://github.com/apache/cloudstack/pull/14038#discussion_r4325729436
##########
engine/orchestration/src/main/java/org/apache/cloudstack/engine/orchestration/NetworkOrchestrator.java:
##########
@@ -1120,10 +1120,12 @@ private NicVO checkForRaceAndAllocateNic(final
NicProfile requested, final Netwo
}
if (vo == null) {
- if (requested.getRequestedIPv4() != null) {
- throw new
InsufficientVirtualNetworkCapacityException("Unable to acquire requested Guest
IP address " + requested.getRequestedIPv4() + " for network " + network,
DataCenter.class, dcVo.getId());
- } else {
- requested.setIPv4Address(null);
+ if (requested != null) {
Review Comment:
Good question. The null `requested` comes from `allocate()`: when a network
is handed in with an empty profile list, the loop adds a single null entry and
calls `allocateNic(null, ...)`. You're right that normal user-vm deploys build
a NicProfile and don't reach it. The reachable path is `reallocate()` in Basic
zones (it puts the network with an empty list) and the other internal
add-network flows that pass no profile. Before the fix, losing the allocation
race there dereferenced the null `requested` and threw NPE instead of retrying;
the guard makes it retry the same as the non-null case.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
