Bug 41831
| Summary: | "mount -a" double mounts mounted partitions | ||
|---|---|---|---|
| Product: | [Retired] Red Hat Linux | Reporter: | greg hosler <greg> |
| Component: | mount | Assignee: | Elliot Lee <sopwith> |
| Status: | CLOSED CURRENTRELEASE | QA Contact: | David Lawrence <dkl> |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | 7.1 | CC: | bugs.michael, mharris |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | i386 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2003-08-01 14:34:18 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
Just from a layman's perspective, w/o looking at code, it would appear as though mount tries to mount a filesystem _before_ checking mtab. Example: (/dev/sdb5 is already mounted on /boot here). # mount /boot I/O error: dev 08:21, sector 2 I/O error: dev 08:21, sector 0 I/O error: dev 08:31, sector 2 I/O error: dev 08:31, sector 0 mount: /dev/sdb5 already mounted or /boot busy mount: according to mtab, /dev/sdb5 is already mounted on /boot -------------------------------------------------- Also, when doing the same with an NFS partition, no errors occur, and 'df' shows the NFS mount being doubly mounted. [Problaby related, although this doesn't match the bug's subject line:] When mounting via the loop device, it is possible to re-use existing mount points, too: mount file1.img /mnt -o loop mount file2.img /mnt -o loop The still mounted virtual fs file1.img gets buried beneath file2.img without any warning or error. When shutting down, only one of the two mounted vfs gets unmounted. If one of the two files is located on a mounted hdd partition, that partition is still in use and hence is not unmounted either, which forces an fsck at next boot time. I would wager this issue doesn't occur anymore. |
From Bugzilla Helper: User-Agent: Mozilla/4.76 [en] (X11; U; Linux 2.4.3-2.9.2smp i686) Description of problem: I think this is a bug, or at least an undesireable side effect to an enhancement. pick one or more mount points that are not presently inuse, and umount them. Then, being lazy, instead of individually mounting the one(s) you just umounted, do: mount -a prior to seawolf, this would mount just the umounted partitions. as of seawolf, this mounts the umounted partitions, and double mounts the mounted ones as well. and since the partitions are already mounted, it will mount the overmount "ro", which can have quite negative impact on a running system (to say the least). Worse still, when you shutdown the system, the double mounted partitions will not be cleanly umounted, which will force a fsck upon the next reboot. This sounds to me like a bug, but I know that mount was changed recently to specifically allow overmounts. just the same, it seems to me that "mount -a" shouldn't double mount a mounted partition. How reproducible: Always Steps to Reproduce: 1.see above 2. 3. Actual Results: mounted partitions are doubly mounted. Expected Results: only partitions in fstab that are not mounted should be mounted. Additional info: