From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.8) Gecko/20050512 Red Hat/1.7.8-1.1.3.1 Description of problem: RHEL3 Kickstart documentation indicates that '--noformat' may be used with the 'raid' directive to prevent MD devices from being formatted. However, issuing 'raid /data --level 0 --device md0 --noformat' fails on hosts with known intact level 0 software RAID arrays with the following error message: "Unable to locate raid device None to use for /data." This seems to be due to the fact that the constructor for the RaidRequestSpec class in partRequests.py is not setting self.device. The attached patch appears to fix the problem. Version-Release number of selected component (if applicable): anaconda-9.1.x How reproducible: Always Steps to Reproduce: 1. Use Kickstart to install a system with an existing software RAID array. 2. Issue a 'raid' directive in your Kickstart configuration with the '--noformat' option. Actual Results: Anaconda displays an "Unable to locate raid device None" message. Expected Results: Anaconda should have found the specified MD device. Additional info:
Created attachment 115093 [details] Patch Which Appears to Correct This Issue
This patch is going to have other side effects that aren't what is wanted (we should only ever set request.device in autopart.py when going through and verifying things). Can you provide /tmp/anaconda.log and /tmp/syslog?
Attached is a patch which causes request.device to be set in the kickstart.defineRaid() method if '--device' is specified. Would setting request.device here be side-effect free? I would naively think so, since this is essentially what is done in definePartition() when '--onpart' is used. I've tested the patch, and it also resolves this issue in our environment.
Created attachment 115297 [details] Another Patch Which Appears to Resolve This Issue
Is this still a problem in RHEL5?