Some problems with my centos 8 began when i entered "yum update -y" command. After that i started receive Illegal instruction (core dumped) messages when i run yum or any python scripts.
During upgrade i got such errors:
Upgrading : NetworkManager-libnm-1:1.30.0-0.5.el8.x86_64 18/168
Running scriptlet: NetworkManager-libnm-1:1.30.0-0.5.el8.x86_64 18/168
Running scriptlet: NetworkManager-1:1.30.0-0.5.el8.x86_64 19/168
Upgrading : NetworkManager-1:1.30.0-0.5.el8.x86_64 19/168
Running scriptlet: NetworkManager-1:1.30.0-0.5.el8.x86_64 19/168
/var/tmp/rpm-tmp.1WBtNY: line 4: 4581 Illegal instruction (core dumped) firewall-cmd --reload --quiet
Upgrading : elfutils-libs-0.182-3.el8.x86_64 20/168
Upgrading : elfutils-debuginfod-client-0.182-3.el8.x86_64 21/168
Upgrading : python3-syspurpose-1.28.9-1.el8.x86_64 22/168
Running scriptlet: selinux-policy-targeted-3.14.3-59.el8.noarch 58/168
Upgrading : selinux-policy-targeted-3.14.3-59.el8.noarch 58/168
Running scriptlet: selinux-policy-targeted-3.14.3-59.el8.noarch 58/168
Upgrading : subscription-manager-1.28.9-1.el8.x86_64 59/168
Running scriptlet: subscription-manager-1.28.9-1.el8.x86_64 59/168
chmod: cannot access '/etc/pki/entitlement/*.pem': No such file or directory
Upgrading : lvm2-8:2.03.11-0.4.20201222gitb84a992.el8.x86_64 60/168
Running scriptlet: lvm2-8:2.03.11-0.4.20201222gitb84a992.el8.x86_64 60/168
Upgrading : kpartx-0.8.4-7.el8.x86_64 61/168
Upgrading : xfsprogs-5.0.0-7.el8.x86_64 62/168
Cleanup : elfutils-default-yama-scope-0.182-2.el8.noarch 154/168
Cleanup : tzdata-java-2020d-1.el8.noarch 155/168
Cleanup : samba-common-4.13.2-5.el8.noarch 156/168
Running scriptlet: NetworkManager-1:1.30.0-0.4.el8.x86_64 157/168
Cleanup : NetworkManager-1:1.30.0-0.4.el8.x86_64 157/168
Running scriptlet: NetworkManager-1:1.30.0-0.4.el8.x86_64 157/168
/var/tmp/rpm-tmp.M7HXpR: line 4: 9248 Illegal instruction (core dumped) firewall-cmd --reload --quiet
Cleanup : NetworkManager-libnm-1:1.30.0-0.4.el8.x86_64 158/168
Running scriptlet: NetworkManager-libnm-1:1.30.0-0.4.el8.x86_64 158/168
Cleanup : perl-libs-4:5.26.3-417.el8.x86_64 159/168
Cleanup : python3-libs-3.6.8-33.el8.x86_64 164/168
Cleanup : glibc-2.28-141.el8.x86_64 165/168
Cleanup : glibc-langpack-en-2.28-141.el8.x86_64 166/168
Cleanup : glibc-common-2.28-141.el8.x86_64 167/168
Cleanup : tzdata-2020d-1.el8.noarch 168/168
Running scriptlet: libwbclient-4.13.3-1.el8.x86_64 168/168
Running scriptlet: tuned-2.15.0-1.el8.noarch 168/168
Running scriptlet: kmod-kvdo-6.2.4.26-76.el8.x86_64 168/168
sort: fflush failed: 'standard output': Broken pipe
sort: write error
gzip: stdout: Broken pipe
gzip: stdout: Broken pipe
sort: write failed: 'standard output': Broken pipe
sort: write error
sort: fflush failed: 'standard output': Broken pipe
sort: write error
gzip: stdout: Broken pipe
gzip: stdout: Broken pipe
sort: write failed: 'standard output': Broken pipe
sort: write error
Running scriptlet: tzdata-2020d-1.el8.noarch 168/168
[/usr/lib/tmpfiles.d/pesign.conf:1] Line references path below legacy directory /var/run/, updating /var/run/pesign вЖТ /run/pesign; please update the tmpfiles.d/ drop-in file accordingly.
Running scriptlet: glibc-common-2.28-145.el8.x86_64 168/168
Verifying : python3-pexpect-4.3.1-3.el8.noarch 1/168
Verifying : python3-ptyprocess-0.5.2-4.el8.noarch 2/168
Verifying : abattis-cantarell-fonts-0.0.25-6.el8.noarch 3/168
Now problems are with yum or with import of some python packages. But curl is ok.
[root@ddns-ard-2 ~]#
[root@ddns-ard-2 ~]# yum --version
Illegal instruction (core dumped)
[root@ddns-ard-2 ~]# yum upgrade
Illegal instruction (core dumped)
[root@ddns-ard-2 ~]# python3
Python 3.6.8 (default, Dec 22 2020, 19:04:08)
[GCC 8.4.1 20200928 (Red Hat 8.4.1-1)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import sys
>>> import pymongo
Illegal instruction (core dumped)
[root@ddns-ard-2 ~]# rpm -q curl
curl-7.61.1-17.el8.x86_64
[root@ddns-ard-2 ~]# curl https://www.google.com | more
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0<!doctype html>
Messages from my /var/log/messages
Jan 12 02:18:02 ddns-ard-2 kernel: traps: yum[3731] trap invalid opcode ip:7efdef6dd59f sp:7ffef19166c0 error:0 in libm-2.28.so[7efdef655000+181000]
Jan 12 02:18:02 ddns-ard-2 systemd[1]: Started Process Core Dump (PID 3732/UID 0).
Jan 12 02:18:02 ddns-ard-2 systemd-coredump[3733]: Resource limits disable core dumping for process 3731 (yum).
Jan 12 02:18:02 ddns-ard-2 systemd-coredump[3733]: Process 3731 (yum) of user 0 dumped core.
Jan 12 02:18:02 ddns-ard-2 systemd[1]: systemd-coredump@278-3732-0.service: Succeeded.
Workaround i founded in internet not helped
[root@ddns-ard-2 ~]# export NSS_DISABLE_HW_AES=1
[root@ddns-ard-2 ~]# yum --version
Illegal instruction (core dumped)
[root@ddns-ard-2 ~]# export NSS_DISABLE_HW_GCM=1
[root@ddns-ard-2 ~]# yum --version
Illegal instruction (core dumped)
[root@ddns-ard-2 ~]#
[root@ddns-ard-2 ~]# dnf
Illegal instruction (core dumped)
[root@ddns-ard-2 ~]#
Any ideas how to resolve this problem ?
I have 5 virtual CPUs with the same cores here is information about one of them:
[root@ddns-ard-2 ~]# cat /proc/cpuinfo
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 63
model name : Intel(R) Xeon(R) CPU E5-2699 v3 @ 2.30GHz
stepping : 2
microcode : 0x38
cpu MHz : 2294.686
cache size : 46080 KB
physical id : 0
siblings : 1
core id : 0
cpu cores : 1
apicid : 0
initial apicid : 0
fpu : yes
fpu_exception : yes
cpuid level : 15
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts mmx fxsr sse sse2 ss syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts nopl xtopology tsc_reliable nonstop_tsc cpuid aperfmperf pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx hypervisor lahf_lm epb pti cqm_llc cqm_occup_llc dtherm ida arat pln pts
bugs : cpu_meltdown spectre_v1 spectre_v2 spec_store_bypass l1tf mds swapgs itlb_multihit
bogomips : 4589.37
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management:
This happens if the hypervisor masks XSAVE support, but does not mask the FMA feature bit in CPUID. It's a bit of a weird hypervisor configuration, which is why it was not caught in testing. Selecting a different CPU model in the hypervisor should work around this (something that supports AVX2 should do it). As a side effect, you will likely get better performance from your VMs.
Some reported this through appropriate channels, and updates have been released: