This post is more than 5 years old
7 Posts
0
6314
November 19th, 2017 04:00
ScaleIO 2.0.1.4 + Centos7.4 SDC
Hi everyone,
I've recently added my first CentOS 7.4 SDC to one of my ScaleIO clusters. However, as soon as I map a volume to the SDC, it freezes, goes 100% CPU and just plain dies.
Have anyone seen this issue before?
I have tested this on both SDC/SDS version 2.0.1.3 and 2.0.1.4.
No Events found!
RHasleton1
73 Posts
1
November 21st, 2017 15:00
This is currently a known issue, but specifically when dealing with VMs. We have seen this behavior with both VMware VMs and Hyper-V guests as well. In this situation, the workaround is to load the SDC at the Hypervisor level and carve up LUNs for the guests that way.
This behavior does not happen with the SDC loaded on bare metal hardware. If the above mirrors your scenario, please follow the above procedure to have those guests use the ScaleIO storage. If you are seeing this behavior with bare metal hardware, please open a service request.
There will be some changes made to the ScaleIO support matrix in the new few weeks regarding RHEL/CentOS 7.4 and it's supportability. Essentially, it will state the above info, that it will be supported on bare-metal hardware, but is not supported at the VM level. We are currently tracking down where the issue resides.
Thanks!
pawelw1
306 Posts
0
November 20th, 2017 01:00
Hi,
Can you please let us know the exact kernel version you use (uname -a) and the version of the SDC client (rpm -qa | grep -i sdc)?
Thank you,
Pawel
munklarsen
7 Posts
0
November 20th, 2017 08:00
Hi Pawel,
Kernel version: Linux 3.10.0-693.5.2.el7.x86_64 #1 SMP Fri Oct 20 20:32:50 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux
SDC Version: EMC-ScaleIO-sdc-2.0-14000.231.el7.x86_64
gabgab1
2 Posts
0
November 20th, 2017 08:00
Hi,
Got exactly the same pb. Here is the kernel and sdc versions:
[root@scaleio01 ~]# rpm -qa | grep -i sdc
EMC-ScaleIO-sdc-2.0-13000.211.el7.x86_64
[root@scaleio01 ~]# uname -a
Linux scaleio01.xxxx.xxx 3.10.0-693.5.2.el7.x86_64 #1 SMP Fri Oct 20 20:32:50 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux
[root@scaleio01 ~]# cat /etc/redhat-release
CentOS Linux release 7.4.1708 (Core)
pawelw1
306 Posts
0
November 21st, 2017 00:00
Thank you,
Let me look into it and get back to you.
Cheers,
Pawel
pawelw1
306 Posts
0
November 21st, 2017 03:00
Hi,
In your case you are using SDC 2.0.1.3 which is not supported with RHEL 7.4; can you please update it to 2.0.1.4 and see if it makes any difference?
Thanks,
Pawel
gabgab1
2 Posts
0
November 21st, 2017 05:00
Hi,
Unfortunately, i dont have access to 2.0.1.4
Gab
klosz007
1 Rookie
•
13 Posts
0
December 27th, 2017 15:00
This is just to let you know that I have hit the same issue - RHEL 7.4 (on VM) + SDC 2.0.1.3 = instant VM freeze and CPUs at 100% after SIO LUN is presented to SDC.
klosz007
1 Rookie
•
13 Posts
0
December 29th, 2017 11:00
The issue seems to be applicable to RHEL 7.4 on VMs + SDC 2.1.0.3/2.1.0.4 only. RHEL 7.3 on VM works fine with SDC 2.1.0.3/2.1.0.4. As long as you don't upgrade via 'yum update' to RHEL 7.4, everything's fine.
Surprisingly SIO 2.1.0.3/2.1.0.4 backend (gateway, MDM, SDC) has no trouble at all when running on RHEL 7.4 VMs.
lucasyvas
1 Message
0
January 2nd, 2018 16:00
This happens on CentOS 7.4 as well with out of the box kernel. 2.0.1.4 does not help the situation and the response from RickH is correct. It will not work for sure on VMWare.
oleginishev
2 Posts
0
January 11th, 2018 02:00
Hello,
For information.
I installed ScaleIO [EMC-ScaleIO-sdc-2.0-14000.231.el7.x86_64.rpm] on CentOS7.4 [kernel 3.10.0-693], this is HyperV VM, and when I map the drive to SDC, I have "NMI watchdog: BUG: soft lock - CPU # 1 stuck for "
But if i downgrade kernel version to [3.10.0-514], i can succesful map drive to SDC.
Can someone comment on this?
klosz007
1 Rookie
•
13 Posts
0
January 11th, 2018 06:00
This is exactly the case being described in this thread.
When SDC 2.0.1.3 or 2.0.1.4 is running on a VM (VMWare or Hyper-V) with RHEL/CentOS 7.4 then the VM will hang after presenting volume to SDC.
Downgrade to RHEL/CentOS 7.3 (that worked for me) or (as advised above) install SDC 2.0.1.3/2.0.1.4 on RHEL/CentOS 7.4 but on bare metal/physical machine.
oleginishev
2 Posts
0
January 12th, 2018 01:00
Thanks!
We updated the kernel to 4.10.5-1.el7.elrepo.x86_64 and successfully mapped drive
vdenaropapa
1 Message
0
January 24th, 2018 07:00
Hello there.
I've got the same issue with this setup:
VMWare Esxi 6.5
Centos 7
Kernel 3.10.0-693.5.2.el7.x86_64
ScaleIO EMC-ScaleIO-sdc-2.0-13000.211.el7.x86_64.rpm
I solved downgrading to kernel 3.10.0-514.el7.x86_64
Thanks
munklarsen
7 Posts
0
February 23rd, 2018 05:00
So it's still an issue with ScaleIO 2.5.0 and Centos 7.4 running 3.10.0-693.11.6.el7.x86_64. Amazing how this issue can be ignored for so long.