Kernel Panic "BUG: unable to handle kernel NULL pointer dereference at 0000000000000010" because of "runc:[2:INIT]" #1857
Comments
|
It looks like something in SELinux is triggering this -- in particular it looks like static void selinux_write_opts(struct seq_file *m,
struct security_mnt_opts *opts)
{
int i;
char *prefix;
for (i = 0; i < opts->num_mnt_opts; i++) {
char *has_comma;
if (opts->mnt_opts[i])
has_comma = strchr(opts->mnt_opts[i], ','); /* <<< panic is here <<< */
else
has_comma = NULL;
switch (opts->mnt_opts_flags[i]) {
case CONTEXT_MNT:
prefix = CONTEXT_STR;
break;
case FSCONTEXT_MNT:
prefix = FSCONTEXT_STR;
break;
case ROOTCONTEXT_MNT:
prefix = ROOTCONTEXT_STR;
break;
case DEFCONTEXT_MNT:
prefix = DEFCONTEXT_STR;
break;
case SBLABEL_MNT:
seq_putc(m, ',');
seq_puts(m, LABELSUPP_STR);
continue;
default:
BUG();
return;
};
/* we need a comma before each option */
seq_putc(m, ',');
seq_puts(m, prefix);
if (has_comma)
seq_putc(m, '\"');
seq_escape(m, opts->mnt_opts[i], "\"\n\\");
if (has_comma)
seq_putc(m, '\"');
}
}Where the line marked is where the panic appears to be. /cc @rhatdan |
|
@cyphar |
|
But to be honest that run (without SELinux) we couldn't catch. We have had bad dump and I not completely sure what exactly happens. |
How did you disable it? |
|
@cyphar It was changed in config and rebooted. |
|
Hi, is the problem solved? I encountered the same bug as well, and the kernel version is also 3.10.0-862.9.1.el7.x86_64 |
|
This should be reported to the (distro) kernel maintainers. If only RHEL 7 users are hitting this, this should be reported to Red Hat. |
Hi.
We worked on Centos 7.4-1708 (kernel 3.10.0-693.11.6) and Docker CE 17.12.0.ce-1.
After system crash we reinstalled OS to Centos 7.5-1804 (kernel 3.10.0-862.9.1) and Docker 18.06.0.ce-3.
Because of that system crash (human factor btw) we started to reindexing ElasticSearch shards (a lot of shards).
And in this circumstances we have had this bug:
SYS
Backtrace
LOG
Situation is very similar to this in case of reproducing the problem.
After all we just downgraded our kernel and docker to last known working configuration (kernel 3.10.0-693.11.6 and Docker CE 17.12.0.ce-1) and all works fine.
#1725
The text was updated successfully, but these errors were encountered: