DHIS version 2.35 is released

I can see in documentation that this feature was added in 2.34 but we came from 2.32, and this is not mentioned in Upgrade Notes in either 2.34 or 2.35.

@Junaid_Mufti

Yes, it probably should be in the upgrade notes. We can fix that.

On the other hand, it should be able to operate at scale. Hopefully disabling it can help to avoid your problem, but it should be safe to have it enabled, and we will need to fix the underlying problem if it is not capable of supporting 50+ DE users.

Please let us know if it helps to disable the service.

Thank you Philip that is most helpful we will try WITH ActiveMQ & Disabled audit state and post our findings.

@Junaid_Mufti Just to be clear. Artemis ActiveMQ is embedded into dhis2 and there is no need to explicitly install anything.
You just have to disable the audit service using the dhis.conf parameters as documented in the link shared by Phil.

1 Like

Hello everyone,
@phil could you have a look at this issue as well please:
[DHIS2-9849] - Jira
Cos this is strange.
Thanks

Hi @Ulanbek

I can reproduce the same double forward slash on our play demo instance, but there I can also still access the URL. Have you only had this error on a local installation?

I have checked with the team and they have (very) recently rewritten the function that does this, so it should be fixed in subsequent patches.

Kind regards,
Phil

Phil,
The Audit - DHIS2 docs for 2.35 has stuff like
“data bytea Compressed Json string of the Audit Entity. It is currently in byte array format and not human-readable, but a mechanism will be introduce in 2.34.1”.

Does that mean this mechanism to read the bytea data has not yet been introduced?

Regards
Calle

Well spotted @Calle_Hedberg. I’ll clarify with the team and we will update that table.

In my case we were using container based LXC system for postgres, application instance & a proxy instance. It did not workout for us. We did try multiple combinations of configurations.

Anyways after moving to a flat installation 2.35, it works beautifully.

Hi @Junaid_Mufti,

Great to hear that it is working fine for you now!

Can you confirm if you are also able to operate 2.35.0 with the default audit service enabled? (i.e. it sounds like some limitation of the containers may have been causing the issue?).

Thanks and kind regards,
Phil

Yes audit service is enabled and its running pretty well so far.

Thank you so much for all the help here.

Hi Junaid

It seems unlikely that this is a container specific issue. To my knowledge the artemis is only communicating internally within DHIS2 on a tcp localhost socket. If, as it sounds, it is working for a while and then failing, this seems much more likely to be a resource leak problem (eg file descriptors). This would cause failure in container before it would on “flat” machine so we need to investigate a little urgently as to what is happening rather than assume the problem is solved. I’ll follow up with you separately.

Cheers
Bob

Hi @phil
I have set 2 2.35.1 instances (one on Amazon AWS and one on a Raspberry). Both produce a lot of those lines mentioned above, especially when importing Events (like 2 lines per Event).

 * WARN  2020-12-01T09:03:50,409 AMQ222061: Client connection failed, clearing up resources for session 20e4139c-33b4-11eb-bf1d-0e84b4f1ffb8 (ServerSessionImpl.java [Thread-2 (activemq-netty-threads)])
 * WARN  2020-12-01T09:03:50,409 AMQ222107: Cleared up resources for session 20e4139c-33b4-11eb-bf1d-0e84b4f1ffb8 (ServerSessionImpl.java [Thread-2 (activemq-netty-threads)])

However Audit table looks just fine.

Besides, skipAudit parameter on import POST query doesn’t seem to be working :frowning:

Hi @phil thanks for this! We are very excited about the analytics offerings of 2.35.0, among other things, and are trying very hard to make 2.35 work for us with a very aggressive upgrade strategy.

The two biggest things that are preventing us from doing so at the moment are the inability to save an event when there are multiple category’ed combinations attached to a program ([DHIS2-10065] - Jira) and the lag that we are experience in the capture app with a large program ([DHIS2-10070] - Jira). Really hope these two issues can be sorted out, or we can find a workaround!

If there is any sense of when 2.35.1 is being planned for, that would help us out a lot in making a plan of when we’ll be able to upgrade. We lose our window if we can’t upgrade before February 2020, so just trying to figure out what our best option truly is. Thanks again for all the work!

1 Like

Hi @ehabets,

With regards to the skipAudit, can you confirm that your user has the “Skip data import audit” authority?

Kind regards,
Phil

Hi @Matthew_Boddie,

I try not to make too many promises about exact dates for patch releases, as they can shift for many reasons.

Generally we currently try to release patches for a given version on approximately 2-month intervals.

We are actually aiming to release 2.35.1 towards the end of this week! At this point we have only a couple of issues to close and would typically not include anything new at this point. The next patch, 2.35.2, would then likely not be before late February 2021.

Having said that, I will raise these issues with our Tracker team and see if we can get some quick feedback.

Kind regards,
Phil

2 Likes

Hi @Matthew_Boddie and @phil,

I am seeing that these two tickets (DHIS2-10065 and DHIS2-10070) are causing issues to Matthew.

DHIS2-10070 is already on our list. We will now also prioritise DHIS2-10065.

Thanks for raising the issues Matthew :slight_smile:

3 Likes

@phil WOWWWW re: this week and the upgrade. This is both amazing and also going to be confounding (in the best of ways) on if we decide to go with 2.34 or 2.35 for our upgrade window. Very much appreciate the quick responses and high level support from @christos and @Karoline, among others already! Will wait with baited breath on any outcomes. Let me know if I can help!

Update as of 12-9-2020, it looks like both of our tickets are “Done” and look to be rolled out for 2.35.1! Can’t wait to see this beautiful 2.35.1 available for us to roll out. Thanks again to all for the great work.

2 Likes

Good information thanks for sharing
Telecommunications Specialist

Hi Bob, hope you are doing well.

Do you have any news on this issue ?
We have the same in 2.35.3 each time the continuous analytics job is executed (1 x minute)

* INFO  2021-05-03T14:32:47,134 [Level: INFO, category: CONTINUOUS_ANALYTICS_TABLE, time: Mon May 03 14:32:47 CEST 2021, message: Analytics tables updated: 00:00:00.192] (InMemoryNotifier.java [taskScheduler-8])
 * INFO  2021-05-03T14:32:47,162 'system-process' update org.hisp.dhis.scheduling.JobConfiguration, name: Continuous Analytics, uid: FyMsmfCZ3il (AuditLogUtil.java [taskScheduler-8])
 * WARN  2021-05-03T14:32:47,203 AMQ222061: Client connection failed, clearing up resources for session aa9a03e2-ac0b-11eb-a94b-005056a8f550 (ServerSessionImpl.java [Thread-4 (activemq-netty-threads)])
 * WARN  2021-05-03T14:32:47,203 AMQ222107: Cleared up resources for session aa9a03e2-ac0b-11eb-a94b-005056a8f550 (ServerSessionImpl.java [Thread-4 (activemq-netty-threads)])

Thanks !