View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0014399||CentOS-7||kernel||public||2018-01-22 13:46||2018-06-10 08:34|
|Priority||normal||Severity||block||Reproducibility||have not tried|
|Platform||Supermicro X10DRH-iT||OS||Kernel 3.10.0-693.11.6||OS Version||7.4.1708|
|Target Version||Fixed in Version|
|Summary||0014399: Kernel Deadlock|
|Description||The recent kernel (Version 3.10.0-693.11.6) with revoked microcode updates (active is Intel Microcode Version Ref. 0x3a 2017-01-30 for Haswell EP CPU, with activated PTI) has produced a kernel deadlock situation (10 of 12 threads blocked, 1x Haswell E5-2620 v3 with Hyperthreading) during heavy fileserver usage (NFSv3 and rsyncd).|
The storage is based on mdraid,lvm and ext4 with a HBA Controller (LSI3008 chipset).
Details see in the attached kernel traces.
After this problem occurs, the storage access was blocked and the system had to be rebooted.
This setup was running with this fileserver load for the first time.
The system was running stable with the same kernel and the microcode
version ref. 0x3b (2017-11-17) for several days (12days).
|Steps To Reproduce||not tried to reproduce because the problem is too bad|
|Tags||No tags attached.|
kernel-3.10.0-693.11.6-deadlock.log (27,123 bytes)
The problem is happening randomly since reported in January. Also with Centos 7.5.1804.
I have also KVM/libvirt running and therefore it could be related to this bug https://bugs.centos.org/view.php?id=13887.