General support questions
-
douglastos
- Posts: 15
- Joined: 2019/01/07 09:49:22
-
Contact:
Post
by douglastos » 2019/01/11 12:56:56
I am using CentOS 7 server which has 77% of use and zabbix is alerting me that the partition is growing.
my question is, what can be "cleaned" in this directory!?!
Last edited by
douglastos on 2019/01/11 15:53:01, edited 1 time in total.
Att.
Douglas Ribeiro
Infrastructure Analyst (Sysadmin)
about.me/douglasribeiro
-
maikcat
- Posts: 7
- Joined: 2019/01/11 13:01:58
Post
by maikcat » 2019/01/11 13:04:17
if there are multiple files starting with vmlinuz* then you have old kernels still installed.
you can use rpm -qa | grep ^kern to check which kernel packages are installed then remove them using rpm -e
Regards
Michael.
-
mashiro2004
- Posts: 68
- Joined: 2018/12/08 21:46:22
- Location: Italy
Post
by mashiro2004 » 2019/01/11 15:48:56
douglastos wrote: ↑2019/01/11 12:56:56
I am using CentOS 7 server which has 77% of use and zabbix is alerting me that the partition is growing.
my question is, what can be "cleaned" in this directory!?!
Have you old kernels?
Launch rpm -q kernel
and
ls -lh /boot
-
douglastos
- Posts: 15
- Joined: 2019/01/07 09:49:22
-
Contact:
Post
by douglastos » 2019/01/11 15:52:22
mashiro2004 wrote: ↑2019/01/11 15:48:56
Have you old kernels?
Launch rpm -q kernel
and
ls -lh /boot
is was exactly that, although the ideal is to have two kernel's, I was told in another forum! Thank you for the feedback !!!
follow the link of the other forum:
https://www.vivaolinux.com.br/topico/Ce ... m-20-livre
Att.
Douglas Ribeiro
Infrastructure Analyst (Sysadmin)
about.me/douglasribeiro
-
mashiro2004
- Posts: 68
- Joined: 2018/12/08 21:46:22
- Location: Italy
Post
by mashiro2004 » 2019/01/11 15:57:34
Good, yes, I also always keep at least 2 kernels
-
TrevorH
- Site Admin
- Posts: 33219
- Joined: 2009/09/24 10:40:56
- Location: Brighton, UK
Post
by TrevorH » 2019/01/11 18:43:22
You can set how many to keep by using installonly_limit= in /etc/yum.conf. See man yum.conf
-
TrevorH
- Site Admin
- Posts: 33219
- Joined: 2009/09/24 10:40:56
- Location: Brighton, UK
Post
by TrevorH » 2019/01/12 15:59:02
I just scrolled back and saw the df output for this. Your /boot is showing as 179MB. That's way too small.
The CentOS 6 default size was 500MB and the CentOS 7 /boot size defaults to 1GB.
Since CentOS 7 always installs at least one "rescue" kernel and the initramfs that goes with that is built with 'hostonly=no' it contains all drivers for all possible hardware and is around 55MB in size so the rescue kernel/initramfs want 60MB on their own. That leaves you 120MB for all normal kernels.
There is then a normal kernel and initramfs for each kernel version that's installed and each one of those is 6MB + the size of the initramfs which varies depending on the hardware it's installed on. I checked my systems and the sizes of the initramfs files for normal kernel entries is betwen 19 and 23MB so each kernel uses close to 30MB. That means your 179MB /boot can hold just under 4 kernels before it hits 100% full and the default yum installonly_limit is 5. Since you warn at 77% full and that's showing as 127MB in your post, you only have to have 2 kernels installed before you hit this limit.
Your /boot is too small.
-
douglastos
- Posts: 15
- Joined: 2019/01/07 09:49:22
-
Contact:
Post
by douglastos » 2019/02/22 10:17:35
TrevorH wrote: ↑2019/01/12 15:59:02
...
Your /boot is too small.
I agree with you @TrevorH plus the operator makes the equipment ready and is already in production, we do not have much to do, but a contour solution that is to remove the kernel.
Thank you for the feedback.
Att.
Douglas Ribeiro
Infrastructure Analyst (Sysadmin)
about.me/douglasribeiro