2

We have a standard configuration for all of our machines. Installation is via a script which installs approx 800 packages over a minimum Centos 6 install.

This has worked fine for 5 years already.

We just got a bunch of new machines fitted with the new Areca 1883 RAID controller and now we get "hw bus time out" during the install. After that, the controller resets the bus, all devices go offline, and the install is corrupt and cannot continue.

This happens at exactly the same place in the install every time: just after installing mesa-dri-drivers.

It appears the mesa-dri-drivers package finishes installation. Further, it has no scripts so I'm thinking the problem is in the next package queued for installation.

Is there any way I can find out what the next package is? Trying rpm --test only shows problems and doesn't show any kind of dry run.

  • Did you try examining the install logs? The other VTs/consoles? – Ignacio Vazquez-Abrams Feb 13 '18 at 3:43
0

Because the dependency graph of so many rpm files is very complex, it is pretty much impossible to find out the installation order in advance. The only solution was to disconnect the RAID array altogether and do the same installation on to a single hard drive. That works without crashing Linux and enables you to catch the entire order.

And it isn't the mesa drivers causing the bus timeouts, the problem is still there after removing mesa from the installation. The bus timeouts happen at about the same point (62%) of the install. Likely some compatibility issue between the Areca 1883 RAID Controller and the HGST HUS726060ALE610 HDDs.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.