Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
FS#5984 - Local partitions spontaneously unmount themselves
|
DetailsI'm running a Samba server for a small business. It's on a PII-266 box with 64MB of RAM; old Gateway, nice generic components. The max load I've ever seen is 10% CPU and RAM usage. Three partitions are mounted on /someshare, /someshare/directory1 and /someshare/directory2. All are ext3 partitions on normal ATA/IDE drives. I don't use Unstable software, and the installation's very slim - Samba, sudo, that's about it. A week or two ago I upgraded the OS from Ubuntu to Arch (which I'm using on four other boxes). Since then, an average of once a day the partitons have spontaneously unmounted themselves. Conditions vary - sometimes it's after work when nobody's on, sometimes twice in one day, sometimes there'll be three days without a problem. One common thread is that it's never been caused by me messing around - I don't touch this box if I can avoid it and haven't been logged into the system on any of the occasions, let alone doing anything as root.
I don't think it's a Samba issue - when the drives unmount Samba happily shows the now-empty directory. The drives remount without complaint, and oddly they don't demand a filesystem check due to an unclean unmount. The root and boot partitions remain mounted and the system keeps running. I know this is vague. I welcome any ideas on how to narrow down the issue. My gut says kernel - the issue doesn't occur on my other Arch boxes and upgrading the kernel via pacman-Syu has reduced but not eliminated the problem. |
This task depends upon
I'd say this should probably be closed under "hardware abuse". Sory for the bother, it was just such a weird problem that it might trigger a dev's memory.