Hi Dojo Community,
Any comments/feedback/suggestion to the following issue is much appreciated:
This issue is in the context of DMA 9.6 CU23. Architecture is a DMS made up with a cluster of 6 failover-pairs.
- We are running a CMTS driver with the following feature:
Via CLI, the driver triggers the CMTS to create a backup of its configuration into a text file.
Upon a 'success' CLI response, the config backup file is generated on the CMTS.
The CMTS driver then attempts to transfer the config backup file from the CMTS to a remote shared folder:
CMTS <---sftp transfer---> DMA (CMTS driver) <---SMB file transfer---> Remote file share
Now the issue that we are trying to solve is that when we have two CMTS elements running backup one after the other the first backup always fails while the second one succeeds. This was tested on different CMTS elements and different DMA failover-pairs with the same result.
A closer look at wireshark traces indicates that both elements will use the same TCP session (i.e.: same SMB file transfer) and that the first file transfer stops (before the full file is transferred) when the second file transfer starts. And we are stumped for ideas on how to manage this as each element is oblivious to what the other elements are doing and it does not seem possible to get each element to establish a new SMB session to the same file share hostname.
Currently an automation script is used to trigger these backups to occur in bulk daily using the scheduler. The script will retrieve all CMTS elements with a particular DMAID and initiates backup all at once. There was an attempt to put in a sleep timer between elements however a long enough sleep timer to allow one CMTS to complete before starting the next was not possible due to the reason that it would make the script run for too long and become auto-terminated after ~12mins. There are about 60 CMTS with the same DMAID and each can take potentially take up to 30s-60s to complete resulting in the need for an automation script to run for 1hr.
Appreciate any ideas the community may have.
Since you have a good idea of how long it will take for the script to run if you iterate through all your CMTS, you should be able to override the termination of your script after 12 mins by adding this :
Thanks for pointing this out Wale, its useful to know that we can extend the timeout period of an automation script. We’ll certainly be exploring this option.