Why do I have two /tmp ?

N

NightTripper

Code: 
# df -h
Filesystem                    Size    Used   Avail Capacity  Mounted on
/dev/da0p2                      4G    368M    3.3G    10%    /
devfs                         1.0k    1.0k      0B   100%    /dev
[color="Blue"]/dev/da0p5                      2G     16M    1.8G     1%    /tmp[/color]
/dev/da0p6                    7.9G    5.3G    1.9G    74%    /usr
/dev/da0p7                     88G    348M     81G     0%    /var
zpool1/noaaport-data          399G     55G    343G    14%    /var/noaaport/data
zpool0/noaaport-nbsp-spool     15G    3.3G     12G    21%    /var/noaaport/nbsp/spool
zpool0                         12G     31k     12G     0%    /zpool0
zpool1                        343G     31k    343G     0%    /zpool1
[color="blue"]/dev/md0                       19M     19M   -1.5M   108%    /tmp[/color]
 
Show us your /etc/fstab.

I've noticed that I'm able to accidentally mount a HD partition on a mountpoint where a malloc drive (what your md0 probably is) is already mounted. Yours seems to be the reverse case.
 
Code: 
file "/etc/fstab", 7 lines
# Device        Mountpoint      FStype  Options Dump    Pass#
/dev/da0p2      /               ufs     rw      1       1
/dev/da0p3      none            swap    sw      0       0
# /dev/da0p4    none            swap    sw      0       0
/dev/da0p5      /tmp            ufs     rw      2       2
/dev/da0p6      /usr            ufs     rw      2       2
/dev/da0p7      /var            ufs     rw      2       2
 
I'm trying to install webmin. But it bombs out after a few seconds. Telling me that there is no room on /tmp??

Mark
 
I can't comment on webmin (never touched the stuff), but something is creating /dev/md0 (presumably allocating 20MB of RAM) and mounting it on /tmp.

If not from /etc/fstab this would be done using mdconfig(), either by hand or from a script. You'll have to search for uses of this program to find the culprit.
 
It seems that webmin is creating the extra /tmp. But why would it make it so small as to fail because of no space??

Mark
 
Code: 
noaaport# mdconfig -d -u 0
mdconfig: ioctl(/dev/mdctl): Device busy
noaaport#

How do I kill the memory drive? Other resources on my system are confusing the memory /tmp with the system /tmp

Mark
 
Well, other than searching what's hogging your /tmp, the easiest would probably be to reboot and keep webmin on a tighter leash on your next attempt (did you ask the webmin people or the port maintainer why webmin is messing with /tmp?)
 
did you ask the webmin people or the port maintainer why webmin is messing with /tmp?
Click to expand...
Yep, but the light is off over there?

I dislike rebooting if I don't have to. But like you said. Thius might be one of those times.

Mark
 
Webmin is disabled but the memory drive is still there. And it will not let me umount /dev/md0 it.

Any suggestions?

Mark
 
Maybe you should check /etc/rc.conf. Mine has something like

Code: 
tmpmfs="YES"
tmpsize="512M"
tmpmfs_flags="-m 0 -o async,noatime -S -p 1777"
 
You've (or a script) has mounted a memory device over /tmp.

To identify what memory device is configured:

Code: 
mdconfig -l -v

Unmount the device:

Code: 
umount /tmp

Destroy the memory device:

Code: 
mdconfig -d -u 0
 
It seems I put that CMD in the rc.conf file. I did it as I was configuring a NOAAport server. I was blindly parroting the instructions. I hope I learned from this.

I emailed the creator of that software about this. He told me to increase the size of the memory based tmp file to 512 meg Everything works now.

Mark
 
But the fact it is dynamic causes a little, and I mean little, loss in speed. And speed is what the author is after.

Mark
 
You must log in or register to reply here.
Back
Top