-
Bug
-
Resolution: Done
-
Highest
-
Oxygen-SR2
-
None
-
None
-
Highest
smalleni and SridharG have found numerous new transaction leaks in internal scale testing, which we suspect could cause out of memory issues. We already chased a lot of those a while ago in NETVIRT-1010 (see also http://blog2.vorburger.ch/2017/09/how-to-find-transaction-related-memory.html), fixed everything we found - and now have a lot of new ones...
The attached x3 log files provide details. We will link (or sub-task) more specific issues; let us keep this issue (only) as the overall parent umbrella issue to link everything together to keep track.
- is blocked by
-
GENIUS-102 New OOM due to more TX leaks seen in the nova-boot scenario (interfacemanager)
- Resolved
-
GENIUS-178 Fix new transaction leaks in idmanager
- Resolved
-
NETVIRT-1385 Transaction leak in NatInterfaceStateChangeListener
- Resolved
-
AAA-177 Suspected transaction leak in MDSALDynamicAuthorizationFilter
- Resolved
-
NETVIRT-1291 New transaction leak in netvirt.fibmanager.VrfEntryListener
- Resolved
-
NETVIRT-1379 Transaction leak in org.opendaylight.netvirt.vpnmanager.VpnInterfaceManager
- Resolved
-
NETVIRT-1380 Transaction leak in org.opendaylight.netvirt.natservice.internal.NatRouterInterfaceListener
- Resolved
- relates to
-
GENIUS-176 Convert all of genius to managed transactions (datastore-constrained)
- In Progress