I have observed a behavior during an initial load that I need help understanding.
I have a server running mySQL and symmetric 3.11.4. I have a mobile device which uses the symmetric API to sync data from the server to the mobile. My server is set to do an automatic initial load when a new node is registered. When I register my mobile via the API things work as expected. The initial load request is started and I can see the batches in sym_outgoing_batch being created on the 'reload' channel. That seems to work just fine. But this next part is what I am having problems understanding. It appears that as the server extracts data from tables to be sent to the new node, it is also creating multiple (as in hundreds) of outgoing batches on the 'default' channel. These batches do not have any sql statements contained in them but do list one of the table names in the summary column. The results are that several thousand 'default' channel batches are created to the 160 'reload' channel batches causing a lengthy time for the mobile to retrieve these empty batches that have nothing to apply. Does anyone have any idea what this is caused by? I'm hoping I have something configured incorrectly causing this behavior.
I do see this in the log during the process of the initial load extracting data which seems to indicate batches being created for the default channel. There are no other nodes defined in the system at this point.
Routing channel 'default' for 60 seconds, routedCount=185, dataEventCount=185, startDataId=174, endDataId=413, readCount=240, peekAheadFillCount=1, dataGaps=2
2021-04-23 18:31:41,875 INFO [server01][RouterService][server01-job-18] Done routing for channel 'default' which took 77 seconds
2021-04-23 18:31:41,891 INFO [server01][RouterService][server01-job-18] Router times for channel 'default':
2021-04-23 18:31:41,891 INFO [server01][RouterService][server01-job-18] Routed channel default, batches.insert.count=240, batches.insert.time.ms=38449, batches.noncommon.count=240, batches.update.time.ms=32138, data.enqueue.time.ms=0, data.events.insert.count=240, data.events.insert.time.ms=5566, data.read.total.time.ms=31, data.routed.count=240, total.time.ms=77657
2021-04-23 18:31:41,891 INFO [server01][RouterService][server01-job-18] Routed 240 data events in 77672 ms
Thank you.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The "default" channel is where changes are captured. That channel will wait until the initial load is complete, so it's not slowing anything down. After the initial load is complete, the changes in batches on the "default" channel will be allowed to sync.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I have observed a behavior during an initial load that I need help understanding.
I have a server running mySQL and symmetric 3.11.4. I have a mobile device which uses the symmetric API to sync data from the server to the mobile. My server is set to do an automatic initial load when a new node is registered. When I register my mobile via the API things work as expected. The initial load request is started and I can see the batches in sym_outgoing_batch being created on the 'reload' channel. That seems to work just fine. But this next part is what I am having problems understanding. It appears that as the server extracts data from tables to be sent to the new node, it is also creating multiple (as in hundreds) of outgoing batches on the 'default' channel. These batches do not have any sql statements contained in them but do list one of the table names in the summary column. The results are that several thousand 'default' channel batches are created to the 160 'reload' channel batches causing a lengthy time for the mobile to retrieve these empty batches that have nothing to apply. Does anyone have any idea what this is caused by? I'm hoping I have something configured incorrectly causing this behavior.
I do see this in the log during the process of the initial load extracting data which seems to indicate batches being created for the default channel. There are no other nodes defined in the system at this point.
Routing channel 'default' for 60 seconds, routedCount=185, dataEventCount=185, startDataId=174, endDataId=413, readCount=240, peekAheadFillCount=1, dataGaps=2
2021-04-23 18:31:41,875 INFO [server01] [RouterService] [server01-job-18] Done routing for channel 'default' which took 77 seconds
2021-04-23 18:31:41,891 INFO [server01] [RouterService] [server01-job-18] Router times for channel 'default':
2021-04-23 18:31:41,891 INFO [server01] [RouterService] [server01-job-18] Routed channel default, batches.insert.count=240, batches.insert.time.ms=38449, batches.noncommon.count=240, batches.update.time.ms=32138, data.enqueue.time.ms=0, data.events.insert.count=240, data.events.insert.time.ms=5566, data.read.total.time.ms=31, data.routed.count=240, total.time.ms=77657
2021-04-23 18:31:41,891 INFO [server01] [RouterService] [server01-job-18] Routed 240 data events in 77672 ms
Thank you.
The "default" channel is where changes are captured. That channel will wait until the initial load is complete, so it's not slowing anything down. After the initial load is complete, the changes in batches on the "default" channel will be allowed to sync.