In bug 102006, a user reported a failure loading the XFS plugin and was able to get around it by simply deleting the xfs plugin before running genkernel. I was not so fortunate, so I had to figure out what the problem was. The first way to solve it, from a genkernel point of view, is to specify a log file that is writeable to evms_activate (/var/log does not exist). Specifying: evms_activate --log-file /temp/evms.log Seems to work for me. The deeper issue in evms however, is that the filesystem appears to be returning the wrong error code. EVMS tests for XFS support my running mkfs.xfs and examining the output. If mkfs.xfs doesn't exist, the process should return ENOENT (which it does under my normal boot). But on the initramfs, the failed process returns EINVAL. From the man page: ENOENT The file filename or a script or ELF interpreter does not exist, or a shared library needed for file or interpreter cannot be found. EINVAL An ELF executable had more than one PT_INTERP segment (i.e., tried to name more than one interpreter). So we're clearly getting the wrong errno. However, this is: a) where my experience starts getting a little thin b) where I start falling asleep If someone else could take a look at it from here, that'd be great. Reproducible: Always Steps to Reproduce:
OK. My evms knowledge is pretty much non-existent. Anyone else have any ideas?
I'm not sure that it's really evms' problem. I think the kernel may be returning the wrong errno when evms_activate does a stat on mkfs.xfs. I can dig deeper if need be. I can't be the only one running EVMS of the initramfs. Does anybody else run EVMS and not have any problems?
hmm, well, this might be kernel related. But I don't think that the problem is solely the non-existence of mkfs.xfs because I don't have that one and I need evms starting in initramfs and it works.
Is this even still an issue?
Closing due to no response. Please reopen if this is still an issue and you have more information.