Identify If Any Node Is Unable To Sync

This site is best viewed in a modern browser with JavaScript enabled.

Identify If Any Node Is Unable To Sync

erabhisinha

Is there any way to identify if any node in the peer to peer topology is unable to sync the database?
In our case there are say 4 nodes and we establish P2P topology and they sync information as expected, but sometimes one of the device say terminal 4) goes out of sync and does not pulls and push information from other peers.
Is there any way to identify whether any terminal is syncing information from another peers?

Bernardo

Hello!

You can try the PRAGMA sync_status command

It returns some info about the peers too, if they are connected. But probably it is most useful in the node that presents the problem itself.

You can check the sync_down_status and the sync_up_status values, as well as checking if the node is actually connected to the other nodes.

If the above stats are not sufficient, please inform. In this case we could insert some other kind of information like the last state id from each node so the app can compare if some other node or itself if behind others.

erabhisinha

Hello,

Can you please explain a bit more about sync_down_status and sync_up_status.

Regarding this out of sync behavior we are unable to identify why it would happen. The devices initially are in sync and after they run for couple of hours one of the device goes out of sync and to bring it back to sync, we have to kill the app of all 4 devices and then re run the app to bring all 4 devices on sync. This can be pain for end users of the application.
Can you please recommend how we can resolve this issue to ensure that once the devices are syncing they do not go out of sync.

-To give you more information about the lite sync integration, we are using P2P topology and while we create database connection we check for db_is_ready state from sync_status command:

var status = DBConnection.Database.Connection.ExecuteScalar<string>("pragma sync_status");
if (status.Contains("\"db_is_ready\": true"))
{
break;
}
Only after above is verified we allow application to start using the database and this works for few hours and post that syncing goes down. From this point we have to restart the app on all four devices to get syncing done.

Thank You!

Bernardo

I was not able to reproduce it. Could you pass more details?

How many transactions per minute or per second each node is generating? Is that continuous or sporadic?

How are the connections between the nodes? Each node is having a single connection to each other node?

Please share the URI used on all nodes, if possible. You can edit/mask the IP addresses.

Is each database opening just a single connection to the local database?

erabhisinha

Hi,

To make a note, the connection initially happens successfully and devices sync information between themselves. But when we keep those devices idle for some time (10+ hours) and then one of the device goes out of sync but remaining three works to sync information properly.

I will answer your questions above in points:

Point 1 - The transactions are sporadic and there frequency is not very high. Number of transaction per device per minute remains below 100.

Point 2 and 3 - We follow P2P topology and the connections are as:

Device-1:

file:/data/user/0/com.companyname.acustom.pos.mobileapp.android/files/.local/share/FirePOS.db?node=primary&total_primary_nodes=4&bind=tcp://0.0.0.0:1201

Device-2:

file:/data/user/0/com.companyname.acustom.pos.mobileapp.android/files/.local/share/FirePOS.db?node=primary&total_primary_nodes=4&bind=tcp://0.0.0.0:1202&connect=tcp://192.168.3.240:1201

Device-3:

file:/data/user/0/com.companyname.acustom.pos.mobileapp.android/files/.local/share/FirePOS.db?node=primary&total_primary_nodes=4&bind=tcp://0.0.0.0:1203&connect=tcp://192.168.3.240:1201,tcp://192.168.3.241:1202

Device-4:

file:/data/user/0/com.companyname.acustom.pos.mobileapp.android/files/.local/share/FirePOS.db?node=primary&total_primary_nodes=4&bind=tcp://0.0.0.0:1204&connect=tcp://192.168.3.240:1201,tcp://192.168.3.241:1202,tcp://192.168.3.243:1203

Point 4: On every device we open only one connection to the local database through singleton class.

Thank You!

Bernardo

Do you know which one of these devices stopped synchronizing? Can you identify it by the URI?

If you can reproduce the problem, can you make your app to display or log the result of PRAGMA sync_status? Mainly on the failing node, but it can be useful on the others too.