Bugs re. apparently never closed TransactionChain in ovsdb


Michael Vorburger <vorburger@...>
 

Hello good people of ovsdb,

Would you guys like to have a closer look at and (probably) provide fix for:


Victor Pickard <vpickard@...>
 

Michael,

I will take a look at these 2 bugs.

Vic


On Tue, Aug 29, 2017 at 6:22 PM, Michael Vorburger <vorburger@...> wrote:
Hello good people of ovsdb,

Would you guys like to have a closer look at and (probably) provide fix for:

_______________________________________________
ovsdb-dev mailing list
ovsdb-dev@....org
https://lists.opendaylight.org/mailman/listinfo/ovsdb-dev



Victor Pickard <vpickard@...>
 

I've pushed patches for these 2 bugs. I don't think these are frequently occurring leaks, as these allocations from SouthboundProvider and HwvtepSouthboundProvider init code.



On Wed, Aug 30, 2017 at 1:06 PM, Victor Pickard <vpickard@...> wrote:
Michael,

I will take a look at these 2 bugs.

Vic


On Tue, Aug 29, 2017 at 6:22 PM, Michael Vorburger <vorburger@...> wrote:
Hello good people of ovsdb,

Would you guys like to have a closer look at and (probably) provide fix for:

_______________________________________________
ovsdb-dev mailing list
ovsdb-dev@...rg
https://lists.opendaylight.org/mailman/listinfo/ovsdb-dev




Michael Vorburger <vorburger@...>
 

On Wed, Aug 30, 2017 at 10:08 PM, Victor Pickard <vpickard@...> wrote:
I've pushed patches for these 2 bugs. I don't think these are frequently occurring leaks, as these allocations from SouthboundProvider and HwvtepSouthboundProvider init code.

yeah, you are right, I've no "proof" yet that Bug 9072 or Bug 9073, or any of the few other ones I opened before yesterday, are the real root cause of https://bugs.opendaylight.org/show_bug.cgi?id=9034 - and if it's only x1 TxChain in init() and never more, then they are probably (TBC!) not - fully agreed.

FYI the cunning plan is that I'll now start looking at the feasibility of back patching the Bug 9060 new Tx leak tooling from master to stable/carbon, so that we can run it on environments under load - and find which of the leaks has a high Nx (not just 1x), in the trace:transactions output...

Having said all of this, IMHO we should still fix ANY leaks ASAP anyway - so Tx for patches!
 
On Wed, Aug 30, 2017 at 1:06 PM, Victor Pickard <vpickard@...> wrote:
Michael,

I will take a look at these 2 bugs.

Vic


On Tue, Aug 29, 2017 at 6:22 PM, Michael Vorburger <vorburger@...> wrote:
Hello good people of ovsdb,

Would you guys like to have a closer look at and (probably) provide fix for:

_______________________________________________
ovsdb-dev mailing list
ovsdb-dev@...rg
https://lists.opendaylight.org/mailman/listinfo/ovsdb-dev





Michael Vorburger <vorburger@...>
 

Hello again good people of ovsdb,

On Thu, Aug 31, 2017 at 11:37 AM, Michael Vorburger <vorburger@...> wrote:
On Wed, Aug 30, 2017 at 10:08 PM, Victor Pickard <vpickard@...> wrote:
I've pushed patches for these 2 bugs. I don't think these are frequently occurring leaks, as these allocations from SouthboundProvider and HwvtepSouthboundProvider init code.

yeah, you are right, I've no "proof" yet that Bug 9072 or Bug 9073, or any of the few other ones I opened before yesterday, are the real root cause of https://bugs.opendaylight.org/show_bug.cgi?id=9034 - and if it's only x1 TxChain in init() and never more, then they are probably (TBC!) not - fully agreed.

FYI the cunning plan is that I'll now start looking at the feasibility of back patching the Bug 9060 new Tx leak tooling from master to stable/carbon, so that we can run it on environments under load - and find which of the leaks has a high Nx (not just 1x), in the trace:transactions output...

Having said all of this, IMHO we should still fix ANY leaks ASAP anyway - so Tx for patches!

we've found two new suspected ReadOnlyTransaction leaks which you guys may would like to have a closer look at?

https://bugs.opendaylight.org/show_bug.cgi?id=9098 OvsdbDataTreeChangeListener

https://bugs.opendaylight.org/show_bug.cgi?id=9097 BridgeConfigReconciliationTask

There two are certainly NOT the cause for https://bugs.opendaylight.org/show_bug.cgi?id=9034 (because they are re. open
ReadOnlyTransaction and not TransactionChain) - but it probably still like a Good Idea to look into and fixes these?


On Wed, Aug 30, 2017 at 1:06 PM, Victor Pickard <vpickard@...> wrote:
Michael,

I will take a look at these 2 bugs.

Vic


On Tue, Aug 29, 2017 at 6:22 PM, Michael Vorburger <vorburger@...> wrote:
Hello good people of ovsdb,

Would you guys like to have a closer look at and (probably) provide fix for:

_______________________________________________
ovsdb-dev mailing list
ovsdb-dev@...rg
https://lists.opendaylight.org/mailman/listinfo/ovsdb-dev