low1 rebooted

[Predrag Punosevac]

Robert MacLachlan <robmacl at cmu.edu> wrote:

> Below is the df on low1.  I guess it's bad that / only has what, 1.2g
> available?   Far from an expert on such things, but the stuff I see on /tmp
> mostly looks like normal crap.  I myself had been directly sticking files
> that could amount to several gig.  I thought that /tmp was stored in memory
> these days?

No /tmp is not in RAM. I ended up rebooting machine to clear the stuff 
as it is not easy even to locate of the cashed files.

Predrag

P.S. You MATLAB should now work as well. I just restared licensing
server.

root at low1$ df -h
Filesystem                  Size  Used Avail Use% Mounted on
/dev/mapper/sl-root          50G   31G   20G  62% /
devtmpfs                    252G     0  252G   0% /dev
tmpfs                       252G     0  252G   0% /dev/shm
tmpfs                       252G  8.7M  252G   1% /run
tmpfs                       252G     0  252G   0% /sys/fs/cgroup
/dev/sda1                   497M  225M  272M  46% /boot
/dev/mapper/sl-home         412G  361G   51G  88% /home
gaia:/mnt/zfsauton/project  1.6T  1.6T   52G  97% /zfsauton/project
gaia:/mnt/zfsauton/zdata    1.4T  1.3T   52G  97% /zfsauton/data
gaia:/mnt/zfsauton/home      18T   16T  2.4T  87% /zfsauton/home
tmpfs                        51G     0   51G   0% /run/user/0





> 
> ram at low1$ df
> Filesystem
> 1K-blocks        Used  Available Use% Mounted on
> /dev/mapper/sl-root                                          52403200
> 51127924    1275276  98% /
> devtmpfs
> 264051888           0  264051888   0% /dev
> tmpfs
> 264096196           0  264096196   0% /dev/shm
> tmpfs                                                       264096196
> 4340860  259755336   2% /run
> tmpfs
> 264096196           0  264096196   0% /sys/fs/cgroup
> /dev/mapper/sl-home                                         430975380
> 377572380   53403000  88% /home
> /dev/sda1                                                      508588
> 211848     296740  42% /boot
> gaia:/mnt/zfsauton/zdata                                   1420997056
> 1366599296   54397760  97% /zfsauton/data
> gaia:/mnt/zfsauton/project                                 1677254720
> 1622856960   54397760  97% /zfsauton/project
> gaia:/mnt/zfsauton/home                                   18837910912
> 16169412288 2668498624  86% /zfsauton/home
> tmpfs
> 52819240           0   52819240   0% /run/user/1101
> tmpfs
> 52819240           0   52819240   0% /run/user/0
> tmpfs
> 52819240           0   52819240   0% /run/user/1009
> tmpfs
> 52819240           0   52819240   0% /run/user/1113
> tmpfs
> 52819240           4   52819236   1% /run/user/1027
> tmpfs
> 52819240           0   52819240   0% /run/user/1117
> tmpfs
> 52819240           0   52819240   0% /run/user/1122
> tmpfs
> 52819240           0   52819240   0% /run/user/1087
> tmpfs
> 52819240           0   52819240   0% /run/user/1031
> tmpfs
> 52819240           0   52819240   0% /run/user/1006
> tmpfs
> 52819240           0   52819240   0% /run/user/1025
> tmpfs
> 52819240           0   52819240   0% /run/user/1057
> tmpfs
> 52819240           0   52819240   0% /run/user/1132
> tmpfs
> 52819240           0   52819240   0% /run/user/1026
> ram at 127.0.0.1:/cygdrive/C/Users/ram/X2GO~1/S-E053~1/spool   732060668
> 388476972  343583696  54%
> /tmp/.x2go-ram/spool/C-ram-53-1519683617_stRTERMINAL_dp32
> ram at low1$
> 
> 
> On Mon, Feb 26, 2018 at 5:48 PM, Predrag Punosevac <predragp at andrew.cmu.edu>
> wrote:
> 
> > Robert MacLachlan <robmacl at cmu.edu> wrote:
> >
> > /zfsauton is over 80% full which is super dangerous but you seems to
> > have a problem with the root system full on one of the machines.
> > Somebody is cashing some crap on /tmp in order to "speed up" something.
> > That is the easiest way to fill up the root directory. Please send an
> > e-mail to users at autonlab and ask people to clear up the cash.
> >
> > Predrag
> >
> >
> > > But it seems to re-occurr.  Is there some limit to the number of files or
> > > amount of data I can write on tmp?  I've noticed that matlab is helpfully
> > > creating 7760 files under /tmp/.felix for some purpose or other.  An
> > then I
> > > am writing some actual data there, but this was only ~600 meg at the time
> > > it's crapping out (total size is a few gig).  A similar run was working,
> > > but maybe I just tipped over some threshold.
> > >
> > >   Rob
> > >
> > >
> > > On Mon, Feb 26, 2018 at 4:15 PM, Robert MacLachlan <robmacl at cmu.edu>
> > wrote:
> > >
> > > > Maybe I was confused.  I was getting out of space error, and I guessed
> > it
> > > > was was on zfsauton because of the minimal free listed..  The error I'm
> > > > getting now on low1 says out of space on /tmp.  How is that even
> > possible
> > > > with tmpfs?  Is there a quota?  I was able to recover by deleting a
> > whole
> > > > bunch of files that I owned in /tmp.
> > > >
> > > >   Rob
> > > >
> > > >
> > > > On Mon, Feb 26, 2018 at 2:18 PM, Predrag Punosevac <
> > > > predragp at andrew.cmu.edu> wrote:
> > > >
> > > >> Robert MacLachlan <robmacl at cmu.edu> wrote:
> > > >>
> > > >> > Just in case nobody else has noticed yet... :)  I have of course
> > been
> > > >> > writing crap, but just a few gig here and there.
> > > >> >
> > > >> >   Rob
> > > >>
> > > >> I got new hard drives but rebuilding file server will require down
> > time.
> > > >> I was and I am still hoping that it can wait the end of the semester.
> > > >> Please use scratch.
> > > >>
> > > >> Predrag
> > > >>
> > > >
> > > >
> >