A suggestion I got from MS support was to break my configuration up into multiple mofs and apply using partial configs. Having done so, it seems that DSC is still looking at the sum of all parts and not the individual mofs.
Originally my node mofs maxed out at around 37KB. I've managed to break them into smaller chunks and remove a lot of extraneous comment lines and break them into 3 different partial configurations that generate mofs at ~7KB. Using hostname.meta.mof, I have configured a node to pull from 3 different automation accounts in Azure. If I comment out 1 of the accounts and only pull from 2 (~14-18KB total), the configuration will run fine. If I add in the 3rd and the overall size exceeds ~23KB, I hit the error: Buffer is not supported. Check encoding and length of the buffer
Are Linux nodes really limited to ~23KB of configuration data or am I missing something to process these partial configuration in series rather than all at once? 23KB seems incredibly small. Alternatively, is there a way to raise this buffer limit on Linux OMI as you can on Windows?
A suggestion I got from MS support was to break my configuration up into multiple mofs and apply using partial configs. Having done so, it seems that DSC is still looking at the sum of all parts and not the individual mofs.
Originally my node mofs maxed out at around 37KB. I've managed to break them into smaller chunks and remove a lot of extraneous comment lines and break them into 3 different partial configurations that generate mofs at ~7KB. Using hostname.meta.mof, I have configured a node to pull from 3 different automation accounts in Azure. If I comment out 1 of the accounts and only pull from 2 (~14-18KB total), the configuration will run fine. If I add in the 3rd and the overall size exceeds ~23KB, I hit the error: Buffer is not supported. Check encoding and length of the buffer
Are Linux nodes really limited to ~23KB of configuration data or am I missing something to process these partial configuration in series rather than all at once? 23KB seems incredibly small. Alternatively, is there a way to raise this buffer limit on Linux OMI as you can on Windows?