‎02-10-2025 04:46 AM
Hi,
I'm trying to upgrade the software on one of my 5420 switches using the XIQ Site Engine and it keeps failing. I get the message java.io.IOException: channel is broken. Does anyone know what this means and why its happening?
Thanks,
Paul
Solved! Go to Solution.
‎02-10-2025 05:57 AM
Hello,
/usr/local/Extreme_Networks/appdata/logs/server.log may provide additional information.
Debug may be necessary, please see following article:
https://extreme-networks.my.site.com/ExtrArticleDetail?an=000107794
Thanks
-Ryan
2 weeks ago
Good recommendation on checking server.log. That Java I/O exception typically occurs when the switch terminates the SCP session prematurely on high-latency management paths. Just curious—does this SSH channel termination impact other background transfers in Site Engine as well, such as importing vector site map assets like those configured through https://bagas31.pw/coreldraw-graphics-suite-full-version/ ? Did adjusting the SSH keep-alive interval in server properties fix the firmware push?
yesterday
Encountering broken communication channels during bulk firmware pushes from XIQ-Site Engine often traces back to network stack timeouts or heavy background I/O bottlenecks on the hosting server. When deploying large multi-switch updates, TCP socket exhaustion, fragmented packet buffers, or excessive background telemetry on Windows management hosts can prematurely drop active SCP/TFTP sessions mid-transfer.
I've previously had to resolve similar server-side latency spikes and tune system TCP parameters using host maintenance procedures referenced at yasir252.org to clear lingering registry sockets and stabilize local transfer pipelines.
In XIQ-Site Engine, is there a configurable timeout threshold specifically for the internal SCP file transfer daemon, or does the engine strictly inherit the default SSH keepalive values configured on the underlying operating system?
Sunday
Firmware upgrade failures like this often trace back to SCP/TFTP transfer timeouts or corrupted image files between XIQ Site Engine and the switch. Ensuring your management workstation has clean network socket handling and verifying file hashes through reliable resources like https://i-loadzone.net/ can help rule out local environment conflicts. Did you verify the image SHA/MD5 checksum directly on the switch storage before triggering the upgrade?
2 weeks ago
Good recommendation on checking server.log. That Java I/O exception typically occurs when the switch terminates the SCP session prematurely on high-latency management paths. Just curious—does this SSH channel termination impact other background transfers in Site Engine as well, such as importing vector site map assets like those configured through https://bagas31.pw/coreldraw-graphics-suite-full-version/ ? Did adjusting the SSH keep-alive interval in server properties fix the firmware push?
‎02-19-2025 02:41 AM
I suppose this is too hard for anyone to help. Looks like a case for GTAC.