High priority task nosql was rejected

jmatos

Hi
Now and then I got the message:
"High priority task: com.infiniteautomation.nosql.MangoNoSqlBatchWriteBehindManager$PointDataCleaner was rejected because it is already running."
Is this in any case a message to worry about? What it means?
Thanks in advance,
José

phildunlap

Hi Jose,

That's not anything to worry about it if pops up every now and again. If you're seeing them on a very regular basis, you can try increasing the point clean interval in the NoSQL system settings. If that doesn't help, then it may point to a large issue.

jmatos

Hi Phildunlap,

I'm afraid it does not help... I increase the point interval to 10000ms and the alarm message still appears to much times. The last time I took a look I have more than 300 messages in one hour.

Also - and I report it to you here because I suspect it have something to do to - (I remind you that this in now in real environment in the client facilities):

A) The Mango appears to be "stoped"(?!). The regular database backup that took place at 00:05 was not done but the configuration backup was. I try to login and there was always a time out error. I try the Webmin and I was able to login. Finaly I turn the Odroid off and on again and the login to Mango was OK.

B) Is there a way to compact (maintenance) the database?

C) Is mandatory to do/enable the NoSQL database backup?

Thanks in advance,
José

phildunlap

A) is your NoSQL module up to date? What version of Mango and the NoSQL module is this?

B) No, purging data is the only way to shrink the NoSQL database currently.

C) No, you can configure NoSQL backups on the /system_settings.shtm page in the NoSQL section.

jmatos

Hi Phildunlap,

Answer to your questions:

Version of mangoNoSqlDatabase 1.3.5 - paid
Version of core 2.8.6 build 606 - paid

If an update is needed please send me the files to do it offline (also with the procedure) because the client facilities are on a secure network without access to the Internet.

Last time on the spot I collect more information I think it maybe useful for you to diagnose what could be wrong...

A) It seems that now and then an alarm from a point on the BACnet does not arrive to MANGO. We could not see the alarm on, but then the alarm off appears correctly.

B) The MANGO log presents alarms with erros reading the several BACnet Fire Centrals and then Return To Normal.

C) There are 8 BACnet Fire Centrals with several hundreds points each one. The update period is set to 6 seconds in each one.

D) Rarely MANGO (3 times) presents me the message "TIME OUT".

E) Now and then when I told MANGO to do a Backup the BACnet Server reports the Client (MANGO) is lost. Once Backup finishes, status OK.

Thanks in advance,
Jose

phildunlap

Hi jmatos,

You can download the files you need to do a manual update from our store site, https://store.infiniteautomation.com

B) What errors? Can you provide the exact message?

C) That should be fine, but you can see the poll times on the data source's edit page in the Runtime Status section.

D) Please provide the entire, exact error message. Matthew (BACnet4J's original author and still maintainer) recently did a fair bit of work on BACnet4J and some on the BACnet data source recently, and I know one this that got better was the stability of polling a data source with many devices, some in communication and some not reachable. If you use the latest Mango and BACnet you would be using this code.

E) This could perhaps be remedied by increasing the timeouts on the local device. Which kind of backup? NoSQL, Database or configuration?

jmatos

Hi Phildunlap

In the update, do I need to worry about the licence or its automatically recognized?

B) 'CDI5': Read properties from device at 192.168.1.105:47808 failed: Completed 1 requests. Excepted on: ReadPropertyMultipleRequest[ListOfReadAcc... ... ...

This is by heart and the text goes on and on... present info from all the points over that BACnet FireCentral. I could sent you the exact message the next time I went to the client.

Thanks and have a nice week-end,
Jose

phildunlap

Your license will only be affected if you're changing the major version of the core, so updating from Mango 2 to Mango 3 requires updating your license.

I would encourage you to try the latest BACnet module for Mango 3, as Matthew has made improvements, as I was referring to in my previous post.

It is most likely a device is not available.

jmatos

Is that meaning that I can install the BACnet module for Mango 3 in the 2.8.8 Core?

Thanks,
Jose

JoelHaggar

No, you need to have the latest Mango 3 Core and the BACnet module. There have been several improvements and bug fixes in this so it's work updating. You can contact support@infiniteautomation.com to see if you can qualify for a free upgrade or get a quote on an upgrade for your existing license.

jmatos

Before I step into an upgrade, I would like to know if the BACnet Datasource interrogates (or/and bind) all the points/objects in the correspondent remote device or only the ones that are effectively create/present as DataPoint ?

Thank you very much,
Jose

phildunlap

The data source only subscribes to points it has been configured to subscribe to, and the poll only asks for points that have been configured as well.

jmatos

Hi Phildunlap

Thanks for your quick reply.
So having us now several datasources is for any chance recommended to aggregate all the point in a single one (we now have 8)? If so is there any quick task to import all the points to a single one?

Thank you,
Jose

jmatos

sorry to insist. Do you have any answer to me?
Thank you,
Jose

phildunlap

Hi jmatos,

I was away. Here's a thread that gives instructions on how to move data points between data sources: https://forum.infiniteautomation.com/topic/2055/changing-datasource-of-datapoint/3

jmatos

So, do you recommend to move the 1.800 points existing between 8 data source a single one?

phildunlap

I would expect it to be slightly different, you can try, but the recommendation we made was to update to the latest BACnet in Mango 3.