Package install fails checksum

[N1029676]

Bottom line up front: Packages fail their checksum check and cause the system to hang sometimes. The error is below.

Edit: I'm running 2.1.3-RELEASE (amd64) | built on Thu May 01 15:52:13 EDT 2014 | FreeBSD 8.3-RELEASE-p16

I recently purchased a dedicated device to run pfSense from Hacom http://www.hacom.net/catalog/mars-ii-openbrick-m-d525-firewall-appliance.

I have configured the networks it is attached to, tested throughput, DHCP assignments, and am playing with IPSec (unsuccesfully). It runs fine and has had an uptime of 13 hours straight (remember I just got this).

But when I try to install ANY package through the package install in web configurator the entire system would hang and require a hard reboot when accessed over wireless. Since then I have started accessing the web configurator through the WAN port (from my existing network before I received this). Now I can see the package installation process move further along and complete the download and then the page times out. BUT after a minute I can access it again without a reboot.

Checking the system logs > general section shows that the package fails the checksum check.

kernel: t/tmp/apkg_bind-9.9.5_10-amd64.pbi failed checksum, the archive may be corrupt.\nof bind-9.9.5_10-amd64 failed!\n\nInstallation aborted.Removing package…\nStarting package deletion for bind-9.9.5_10-amd64...done.\nRemoving bind components...\nConfiguration... done.\ndone.\nFailed to install package.\n";

I have googled and searched this forum with little luck. Any help would be appreciated.

[jimp]

Is that a full install or NanoBSD?

If it's NanoBSD it could be filling up the tmp and/or var memory disks causing it to believe something is corrupt.

[N1029676]

It was NanoBSD.

I finally just did a reinstall, configured the basics, and downloaded a package again to test.

I downloaded Arping and the whole process was done in less than 8 seconds. Before it took almost a minute to download that package. I'm not sure what was going on but this was the 'fix'.

Thanks for your time!