After Martin's bug report, not much happened to that RC issue in rescue, which is likely to destroy users' data in some situations. The very naive way to try fixing this is in the attached patch: dropping auto-assembly aof RAID arrays by mdadm. Is this too naive or could it be seen as a valuable safe solution to that problem? --
--- rescue-mode.postinst.old 2008-11-29 18:45:57.385261753 +0100 +++ rescue-mode.postinst 2008-11-29 18:46:11.289261411 +0100 @@ -84,7 +84,7 @@ log-output -t rescue --pass-stdout \ mdadm --examine --scan --config=partitions > /tmp/mdadm.conf || true log-output -t rescue \ - mdadm --assemble --scan --run --config=/tmp/mdadm.conf --auto=yes || true + mdadm --assemble --scan --run --config=/tmp/mdadm.conf || true # LVM support try_load_module dm-mod
Attachment:
signature.asc
Description: Digital signature