No space left on device
punosevac72 at gmail.com
Sat Jul 23 14:05:11 PDT 2022
Matthew Dillon <dillon at backplane.com> wrote:
It is always a privilege to hear from you!
> Since the manual hammer cleanup worked, the question is now whether
> the upgrade actually burned through 9GB between cleanups or whether
> the hammer cleanup is no longer running automatically. You should be
> able to check root mail from the cron jobs to see if it is properly
> running hammer cleanup each day or not. That root filesystem is
> quite small for either H1 or H2.
I checked emails. Everything appears OK. An email received from daily
cron job indicates that /sbin/hammer cleanup is running normally.
According to an email received from weekly
Rebuilding locate database:
was run successfully. I did receive email confirming that /data and
/backup file system setup manually by me was cleaned properly. I will
forward these emails to the the same email address I receive my firewall
and desktop daily reports not to be caught off guard again. That was
stupid of me to sit on 643 emails.
> Other places to look for cleaning stuff up. Check for crash dumps in
> /var/crash that you can delete. And also check for excessively cached
That was one of the first places I checked. I have been bitten by that
snake numerous times at work. You would be surprised how many times
"deep" learning guys can run out of 1TB of RAM a day. Multiply that with
50 machines and you get the idea what I do during my lunch break :)
> 'old' (or even new) package tarballs in /var/cache/pkg. You can run
> 'pkg clean' to clean them out.
Da! Just reclaimed 500MB with pkg clean and additional 300MB with
pkg autoremove. I used to run those two commands after every pkg upgrade
but you guys spoiled me by the stability of the system. I log into this
machine a 5-6 times a year and it is just doing what is supposed to
As a final note this file server was provisioned circa 2016 (Dfly 4.6
release but data drives are from an earlier incarnation 2013 Dfly 3.4)
with two M.2 SSD drives of size 32 GB each (much smaller in reality due
to the actual use of binary system). One of them is used for swapcache
and another one you saw in my original email is a root file system.
Apart of two minor huck-ups never saw a problem. That is a testimony to
your work. Once I go for H2. I keep looking at Michael Neumann's emails
and indication that
turned by default on H2. Once you release 6.4 I will carefully go
through release notes and give H2 a go.
FYI I did have a major HDD failure of a master drive 2-3 years ago. I
promoted slave into the master and added a new slave. Another testimony
to your work.
Most Kind Regards,
More information about the Users