Hi Shuva,
In addition to the recreation efforts with higher debug level, please note that
this issue occurred long after the DPN was connected (hard to see that in the
log as it is trimmed).
Looking in the log Dotan has attached to the ticket, between 2016-11-08
08:28:32 to 2016-11-08 09:22:18, you can see ~250 successful
InterfaceStateChangeListeners events (which I believe are triggered from the
port status messages). At this point something breaks, and you start seeing the
“Error processing port status” WARNs.
I also just noticed that these WARNs have started occurring after the following
messages, could they give us some lead?
2016-11-08 09:22:19,682 | WARN | CommitFutures-4 | TransactionChainManager
| 291 - org.opendaylight.openflowplugin.impl - 0.3.1.SNAPSHOT | txChain
failed -> recreating due to {}
2016-11-08 09:22:19,684 | ERROR | CommitFutures-4 | ExecutionList
| 38 - com.google.guava - 18.0.0 | RuntimeException while executing
runnable com.google.common.util.concurrent.Futures$6@680785f9 with executor
INSTANCE
2016-11-08 09:22:19,698 | ERROR | CommitFutures-4 | TransactionChainManager
| 291 - org.opendaylight.openflowplugin.impl - 0.3.1.SNAPSHOT |
Transaction commit failed.
Thanks,
Koby
From: Kochba, Alon
Sent: Tuesday, December 6, 2016 1:49 PM
To: Shuva Jyoti Kar <[email protected]>; Vishal Thapar
<[email protected]>; Sam Hague <[email protected]>; Dayavanti Gopal
Kamath <[email protected]>
Cc: Anil Vishnoi <[email protected]>; Aizer, Koby <[email protected]>;
[email protected]
Subject: RE: FW: [openflowplugin-dev] Error processing port status message
Thanks Shuva,
It happened to Dotan with only one compute.
Can you try the script he provided to recreate?
I asked him to try to recreate it with the debug logs and tcpdump in any case.
--alon
From: Shuva Jyoti Kar [mailto:[email protected]]
Sent: Tuesday, 6 December 2016 07:26
To: Kochba, Alon <[email protected]<mailto:[email protected]>>; Vishal Thapar
<[email protected]<mailto:[email protected]>>; Sam Hague
<[email protected]<mailto:[email protected]>>; Dayavanti Gopal Kamath
<[email protected]<mailto:[email protected]>>
Cc: Anil Vishnoi <[email protected]<mailto:[email protected]>>; Aizer,
Koby <[email protected]<mailto:[email protected]>>;
[email protected]<mailto:[email protected]>
Subject: RE: FW: [openflowplugin-dev] Error processing port status message
Hi Alon,
Sorry for the late response, got pulled into multiple things.
The actual issue was reported as a part of bug
6595(https://bugs.opendaylight.org/show_bug.cgi?id=6595) wherein the OpenFlow
plugin failed to process a port status message because the message came from a
switch before it started the service as MASTER for the switch.
Analysing the logs attached to this bug I was not able to come to any
conclusion. To analyse the issue further we need the logs by setting the log
level as
log:set DEBUG org.opendaylight.openflowplugin.impl
I assume that we see the issue when the number of switches(DPNs) are high,
hence we need to check what is the typical delay between the MASTERSHIP grant
and the ports coming up and adjust the delay accordingly.
We can also get the relevant information in a wireshark capture that would
demarcate when the ROLE_REQ is processed by the switch and when the PORT STATUS
is sent.
Thanks,
shuva
From: Kochba, Alon [mailto:[email protected]]
Sent: Monday, December 05, 2016 9:27 PM
To: Vishal Thapar; Sam Hague; Dayavanti Gopal Kamath; Shuva Jyoti Kar
Cc: Anil Vishnoi; Aizer, Koby
Subject: RE: FW: [openflowplugin-dev] Error processing port status message
Hi, still waiting for *any* response.
Please give some feedback as SR2 is very near and this is possibly blocking.
Thanks,
--alon
From: Vishal Thapar [mailto:[email protected]]
Sent: Thursday, 1 December 2016 14:53
To: Kochba, Alon <[email protected]<mailto:[email protected]>>; Sam Hague
<[email protected]<mailto:[email protected]>>; Dayavanti Gopal Kamath
<[email protected]<mailto:[email protected]>>;
Shuva Jyoti Kar
<[email protected]<mailto:[email protected]>>
Cc: Anil Vishnoi <[email protected]<mailto:[email protected]>>; Aizer,
Koby <[email protected]<mailto:[email protected]>>
Subject: RE: FW: [openflowplugin-dev] Error processing port status message
Sory Alon, totally forgot about it
Adding Shuva.
From: Kochba, Alon [mailto:[email protected]]
Sent: 01 December 2016 18:20
To: Vishal Thapar
<[email protected]<mailto:[email protected]>>; Sam Hague
<[email protected]<mailto:[email protected]>>; Dayavanti Gopal Kamath
<[email protected]<mailto:[email protected]>>
Cc: Anil Vishnoi <[email protected]<mailto:[email protected]>>; Aizer,
Koby <[email protected]<mailto:[email protected]>>
Subject: RE: FW: [openflowplugin-dev] Error processing port status message
Hi,
Our QA stepped on this bug again today, could anyone please provide some input?
Thanks,
--alon
From: Vishal Thapar [mailto:[email protected]]
Sent: Monday, 28 November 2016 19:26
To: Sam Hague <[email protected]<mailto:[email protected]>>; Kochba, Alon
<[email protected]<mailto:[email protected]>>; Dayavanti Gopal Kamath
<[email protected]<mailto:[email protected]>>
Cc: Anil Vishnoi <[email protected]<mailto:[email protected]>>; Aizer,
Koby <[email protected]<mailto:[email protected]>>
Subject: RE: FW: [openflowplugin-dev] Error processing port status message
Hi Sam,
Tomorrow we’re having an offsite so no one will be able to take a look till
Wednesday. If Anil is unable to take a quick look by them, I’ll bring it up
with Shuva.
Regarsd,
Vishal.
From: Sam Hague [mailto:[email protected]]
Sent: 28 November 2016 22:53
To: Kochba, Alon <[email protected]<mailto:[email protected]>>; Dayavanti Gopal
Kamath
<[email protected]<mailto:[email protected]>>;
Vishal Thapar <[email protected]<mailto:[email protected]>>
Cc: Anil Vishnoi <[email protected]<mailto:[email protected]>>; Aizer,
Koby <[email protected]<mailto:[email protected]>>
Subject: Re: FW: [openflowplugin-dev] Error processing port status message
Adding Daya and Vishal if they know someone who could take a look at this.
On Sun, Nov 27, 2016 at 8:22 AM, Kochba, Alon
<[email protected]<mailto:[email protected]>> wrote:
Hey guys,
We encountered a possibly critical bug in openflowplugin a few weeks back.
Tried reaching out to openflowplugin-dev multiple times but no answer since the
first one - Anil/Sam could you assist?
It seems to be easily reproducible and causes a critical (if not blocker) state
on the OVS where new ports cannot be created.
I raised the bug severity to critical, was hoping to get confirmation from
openflowplugin first but maybe this is the way.
thanks,
--alon
From:
[email protected]<mailto:[email protected]>
[mailto:[email protected]<mailto:[email protected]>]
On Behalf Of Kochba, Alon
Sent: Tuesday, 22 November 2016 20:30
To: Aizer, Koby <[email protected]<mailto:[email protected]>>;
[email protected]<mailto:[email protected]>
Cc:
[email protected]<mailto:[email protected]>;
Sokolover, Dotan <[email protected]<mailto:[email protected]>>
Subject: Re: [openflowplugin-dev] Error processing port status message
Bumping this, Shuva or others could we get your inputs please?
If confirmed this seems like a critical bug.
Thanks!
--alon
From:
[email protected]<mailto:[email protected]>
[mailto:[email protected]] On Behalf Of Aizer,
Koby
Sent: Wednesday, 16 November 2016 18:29
To: [email protected]<mailto:[email protected]>
Cc:
[email protected]<mailto:[email protected]>;
Sokolover, Dotan <[email protected]<mailto:[email protected]>>
Subject: [openflowplugin-dev] Error processing port status message
Hi Shuva,
(Sorry for creating a new thread, but I was somehow missing the original thread
[1] in my mail client)
We are hitting bug [2] when using netvirt on Boron SR2. This usually happens
when creating/deleting a large number of VMs.
Dotan has added to the bug description a small script that eventually
reproduces it, and he did it while TransactionChainManager’s debug level is
increased.
We didn’t see any DEBUG messages from that class (there were several
ERROR/WARNs though – added the relevant section of the log to the bug).
Please let us know if we can do anything else in order to help getting this
fixed, as it seem quite critical for us (once this issue occurs it persists and
no more port updates are received from the switch).
Thanks,
Koby
[1]
https://lists.opendaylight.org/pipermail/openflowplugin-dev/2016-October/006155.html
[2] https://bugs.opendaylight.org/show_bug.cgi?id=6908
_______________________________________________
openflowplugin-dev mailing list
[email protected]
https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev