Thanks Daniel,

That sound like a good idea to possibly apply it to 18.12 too

Jacques

Le 05/09/2023 à 21:40, Daniel Watford a écrit :
Hello Jacques and all,

A restart policy has been applied to the demo-trunk and demo-next
containers under https://issues.apache.org/jira/browse/OFBIZ-12853.

We'll have to wait for the next VM restart to confirm correct behaviour,
but I did test by stopping and starting the docker service which
successfully brought the container based demo containers back online.

If this change is proven to work, perhaps we should consider making a
docker version of 18.10 too, particularly since 18.10 will remain the
stable OFBiz version for a little while longer.

Thanks,

Dan.

On Tue, 5 Sept 2023 at 10:05, Jacques Le Roux <jacques.le.r...@les7arts.com>
wrote:

Hi,

Same today, when asking for "ps aux | grep ofbiz" no processes at all.

The last information we got for stable demo (18.12 branch) is (UTC time)

2023-09-05 06:40:16,234 |7.0.0.1-18009-exec-9
|LoginWorker                   |W| Received a null Security object from
HttpServletRequest
2023-09-05 06:40:20,157 |7.0.0.1-18009-exec-4
|LoginWorker                   |W| Received a null Security object from
HttpServletRequest
2023-09-05 06:40:29,855 |Catalina-utility-2
|ControlEventListener          |I| Destroying session:  hidden sessionId by
default.
2023-09-05 06:40:59,863 |Catalina-utility-2
|ControlEventListener          |I| Destroying session:  hidden sessionId by
default.
2023-09-05 06:40:59,864 |Catalina-utility-2
|ControlEventListener          |I| Destroying session:  hidden sessionId by
default.
2023-09-05 06:40:59,866 |Catalina-utility-2
|ControlEventListener          |I| Destroying session:  hidden sessionId by
default.
2023-09-05 06:40:59,868 |Catalina-utility-2
|ControlEventListener          |I| Destroying session:  hidden sessionId by
default.
2023-09-05 06:41:29,877 |Catalina-utility-2
|ControlEventListener          |I| Destroying session:  hidden sessionId by
default.
2023-09-05 06:41:59,891 |Catalina-utility-2
|ControlEventListener          |I| Destroying session:  hidden sessionId by
default.
2023-09-05 06:41:59,893 |Catalina-utility-2
|ControlEventListener          |I| Destroying session:  hidden sessionId by
default.
2023-09-05 06:41:59,894 |Catalina-utility-2
|ControlEventListener          |I| Destroying session:  hidden sessionId by
default.
2023-09-05 06:41:59,895 |Catalina-utility-2
|ControlEventListener          |I| Destroying session:  hidden sessionId by
default.
2023-09-05 06:41:59,895 |Catalina-utility-2
|ControlEventListener          |I| Could not find visit value object in
session [ hidden sessionId by default.] that is being destroyed
2023-09-05 06:41:59,895 |Catalina-utility-2
|ControlEventListener          |I| Destroying session:  hidden sessionId by
default.
2023-09-05 06:41:59,896 |Catalina-utility-2
|ControlEventListener          |I| Destroying session:  hidden sessionId by
default.

So I again suspect someone (Infra or the VM provider or someone in the
middle) has restarted the VM.

I restart manually all the demos , wait 5 to 10 mins...

Jacques
Le 31/08/2023 à 17:08, Daniel Watford a écrit :

Hi Eugen,

It's still on my list to address these issues, but I have not been able to
spend time on them yet. If someone else is able to address them in the
meantime, that would be great, but I think access to the host is needed to
apply and test those changes.

 From memory, version 3 didn't support some of the features we are using
such as memory limits - but can't be sure.   There was a problem with using
version 3 which prompted a downgrade to 2.4.

Thanks,

Dan.

On Thu, 31 Aug 2023 at 15:33, <eugen.s...@netdava.com> wrote:

Hi,

It seems that some services have _restart: "no"_  and some are missing
this option.

The docs for restart policy are here:
https://docs.docker.com/compose/compose-file/compose-file-v2/#restart

I did notice the compose version is 2.4. Any idea why not use version 3
(which is closer to swarm stack IMO)?


https://github.com/apache/ofbiz-tools/blob/a20f9f3c750600164d08b54b80865093399a24de/demo-backup/ofbizdocker/home/ofbizdocker/demo-next/docker-compose.yml#L9

On 31.08.2023 13:43, Jacques Le Roux <jacques.le.r...@les7arts.com>
wrote:
Le 31/08/2023 à 11:25, Eugen Stan a écrit :
La 30.08.2023 11:00, Jacques Le Roux a scris:
Hi,

All demos are down (no processes running). I suspect a restart of
the
VM. I restart them all manually, wait 5 to 10 mins...

Jacques

Hello Jacques,

I presume you are talking about docker deployments.
Can I take a look at how they are deployed ?

I keep reading about them being down because of restarts or some
other
reason.
I believe I can help. Docker has options to restart automatically:

https://docs.docker.com/config/containers/start-containers-automatically/
I think adding `unless-stopped` restart policy should do it.
You can also go with `always`.

docker run -d --restart unless-stopped redis

Regards,
Hi Eugen,

Actually, this time at least, Docker is not concerned. When I said (no
processes running) I meant no OFBiz process.
It happens from time to time. I guess this was due by a reported outage
to Infra (even backup). Almost all VMs are externalised now.

For Docker, all the information is there:

https://github.com/apache/ofbiz-tools/tree/master/demo-backup/ofbizdocker
HTH

Jacques



--
Daniel Watford


Reply via email to