[OVSDB-397] operational datastore is missing for node if the ucast macs are high in the node Created: 19/Jan/17  Updated: 01/May/19  Resolved: 01/May/19

Status: Resolved
Project: ovsdb
Component/s: Southbound.hw_vtep
Affects Version/s: unspecified
Fix Version/s: None

Type: Bug
Reporter: suneel verma Assignee: suneel verma
Resolution: Won't Do Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Operating System: All
Platform: All


External issue ID: 7600

 Description   

If the device has more ucast macs (remote+local) say 4k
the monitor hardware table json rpc request is timing out ( default timer being 1sec ) and operational datastore for the node is not getting filled.

Can we just increase the reaper interval timeout from 1 sec ?
Will that have some other side affects ?



 Comments   
Comment by Vishal Thapar [ 23/Jan/17 ]

Suneelu, did this get fixed as part of other/HA changes?

Comment by Anil Vishnoi [ 02/Feb/17 ]

The only side effect it will have is that it will accumulate more future objects in wait state so memory consumption will increase.

I believe 1 second is pretty long duration, but if your controller has enough heap space, increasing the reaper timeout (ovsdb-rpc-task-timeout) should be okay.

Comment by Anil Vishnoi [ 22/Feb/17 ]

Hi Suneel,

Any update on this patch?

Thanks
Anil

Comment by daya kamath [ 07/Mar/17 ]

we should check whether the delay is caused on the switch (ovs-vtep or physical), or in the plugin to do the unmarshalling of the msg.
can we also check if ovsdb protocol itself supports batching of large messages into smaller chunks

Comment by Anil Vishnoi [ 01/May/19 ]

No activity on the bug for 2 years. Closing the bug, please re-open if you plan to work on it again.

Generated at Wed Feb 07 20:36:17 UTC 2024 using Jira 8.20.10#820010-sha1:ace47f9899e9ee25d7157d59aa17ab06aee30d3d.