Risk of deleting important data when RAID fails
Bug #61018 reported by
Guillermo Pérez
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| sysvinit (Ubuntu) |
Fix Released
|
Undecided
|
Unassigned | ||
Bug Description
Binary package hint: initscripts
When RAID fails because of disk failure a boot could alter the order of the mdX partitions. If md4 failed to initialize md5 could become md4.
I have a deletion of the contents of a partition because of this (md4 was /tmp). Fortunately md5 was just /var/log so no very important data was lost but anyway it was frightening just think on the possibility of deleting de /home because of this.
A way of detecting if the partition is the real tmp would be really nice. Perhaps a hidden file /tmp/.tmp_partition or better detecting if there has been any problem with the boot process and skip the deletion.
For now on my /tmp partition will be always the last one.
Revision history for this message
| Dennis Kaarsemaker (dennis) wrote Re: [Bug 61018] Risk of deleting important data when RAID fails | #1 |
Revision history for this message
| Guillermo Pérez (bisho) wrote | #2 |
Revision history for this message
| Dennis Kaarsemaker (dennis) wrote Re: [Bug 61018] Re: Risk of deleting important data when RAID fails | #3 |
Revision history for this message
| Scott James Remnant (Canonical) (canonical-scott) wrote | #4 |
| Changed in sysvinit: | |
| status: | Unconfirmed → Fix Released |
To post a comment you must log in.
This is one of the reasons ubuntu now uses UUID for mounting, and
not /dev/foo