I have upgrade a few systems to DM10 over the past few days and on each on it have be receiving the following notice. This is happening with going from 9.6 to 10.0 and also when going from 10.0 to 10.1. Any ideas why I am getting this?
Hey Miguel,
This notice is a Cube-side notice while Cube is monitoring the progress of an update.
During an upgrade, the upgrade window will be polling for updates through port 8004 (the same port which is typically used for .NET Remoting DataMiner communication).
In the first part of the upgrade, the server side SLNet process can still answer these queries, even when the DataMiner agent is stopped by the upgrade.
The last part of the upgrade will also stop that SLNet process in order to replace the binaries and restart it. During that time, a helper process is launched on the server to still be able to reply to the poll requests that come in via port 8004.
The notice "Failed getting progress from http://x.x.x.x:8004/UpgradeService: Unable to connect to the remote server" means that the upgrade window is unable to request the progress because the server is not responding or not listening on the port.
Possible causes:
- Server is in between closing the port (shutting down the SLNet process) and setting up the listening port from the helper process
- Server really is unreachable. I guess this could happen while upgrading a Failover cluster through its virtual IP address (which gets removed while the DMA is restarting)
Judging from the two agent-progress tab pages in the screenshot and the time of the notice (5 minutes after the upgrade launched), I conclude that progress has been received prior to getting the notice.
I would not assume that anything went wrong with the upgrade itself. If in doubt, check the upgrade.log files in c:\skyline dataminer\upgrades.
A couple of follow-up questions:
1) How is the user supposed to determine if things like these are an indication of real problems of if they can be ignored?
2) Can the user just ignore “Notices” and only be concerned with “Errors”?
3) When upgrading a Failover system via the Cube should the Cube session be connected to the actual IP and NOT the VIP?
4) When the Skyline Test department is doing upgrade testing, what do they do when they encounter notices and/or errors.
From a customer point of view, when performing upgrades or other similar operations, if the operation completes with any indication other than successful, such as these notifications, it adds confusion to the process and I lose confidence that I have a reliable system. Any official additional information or direction on how to interpret these types of items would be a big help. My organization has probably ~20 test and >50 customer systems all getting upgraded many times a year by a support staff of >20, so when things like this happen it generates a lot of white noise as I’m sure you can imagine.
Thanks