Webinar: Red Hat OpenShift for IBM Z and LinuxONE on RHEL 8.3 KVM

Join us for our webinar on Wednesday, April 21, 11:00 AM - 12:00 PM EST!

Abstract

Red Hat OpenShift is available on RHEL 8.3 KVM starting with Red Hat OpenShift version 4.7 on IBM Z and LinuxONE. We discuss the deployment of a Red Hat OpenShift Cluster on RHEL KVM from a high-level perspective, including supported configurations and requirements, especially the available network and storage options.
Furthermore, we explain the installation steps of Red Hat OpenShift 4.7 on RHEL KVM in detail, including best practices and a short excursion on cluster debugging.

Speakers

  • Dr. Wolfgang Voesch, Iteration Manager - OpenShift on IBM Z and LinuxONE
  • Holger Wolf, Product Owner - OpenShift on Linux on IBM Z and LinuxONE

Registration

Register here. You can check the system requirements here.
After registering, you will receive a confirmation email containing information about joining the webinar.

Replay & Archive

All sessions are recorded. For the archive as well as a replay and handout of this session and all previous webinars see here.

New Release: Linux Kernel 5.11

Linux kernel 5.11 (available here) has been released. Here are the highlights from a Linux on Z perspective:

  • Crypto  
    • zcrypt: Applications that use crypto adapters in the early boot phase (e.g. protected key dm-crypt) may occasionally detect that the crypto devices are not (yet?) ready for use which may result in unexpected error situations. So far, the Linux kernel did not indicate when AP bus scanning was complete and all crypto devices online.
      With the new feature, the kernel sends uevents to indicate that the AP bus scan is complete and that the bindings are complete. In addition, two new sysfs attributes are introduced that indicate the completion of the AP bus scan and the progress in device bindings. These mechanisms allow applications (or udev rules) to delay the use of crypto devices until they initialization is complete.
  • Networking
    • SMC-R Link Group Support: Kernel-side support for monitoring link groups. Requires smc-tools v1.4 for userspace-side exploitation.
      In particular, state and protection against failures on a per-link basis are available, so that administrators can identify critical infrastructure and act accordingly.
        $ smcr device show all
        Net-Dev   IB-Dev  IB-P  IB-State  Type          Crit  #Links  PNET-ID
        ens281    mlx4_0  1       ACTIVE  RoCE_EXPRESS  No    1       NET1
        ens281d1  mlx4_0  2     INACTIVE  RoCE
      _EXPRESS  No    1       NET2
      Also provides insights into link groups (further details available with option --detail):
        $ smcr linkgroup link-show all
        LG-ID     LG-Role  LG-Type  Net-Dev   Link-State   #Conns
        00000100  CLNT     SYM      ens281    LINK_ACTIVE  1
        00000100  CLNT     SYM      enP1s282  LINK_ACTIVE  0

  • Storage
    • Fibre Channel Endpoint Security capability: Exposes the Fibre Channel Endpoint Security (FCES) for DASDs via the sysfs filesystem.
        $ cat /sys/bus/ccw/devices/0.0.c600/fc_security
        Encryption

      Further integrated in lsdasd and lszdev commands. See the following output for further details:
        $ chzdev dasd --help-attribute fc_security
        ATTRIBUTE fc_security
        DESCRIPTION
        This read-only attribute shows the Fibre Channel Endpoint Security 
        status of the connection to the
        DASD device:
          Unsupported :   The DASD device does not support Fibre Channel
                          Endpoint Security
          Inconsistent :  The operational channel paths of the DASD device
                          report inconsistent Fibre Channel Endpoint
                          Security status
          Authentication: The connection has been
                          authenticated
          Encryption :    The connection is encrypted

New Release: s390-tools v2.16

s390-tools v2.16 is out. This release is in support of Linux kernels 5.10 and 5.11.

The most notable change is the addition of a new tool called hsci used to control and show HSCI (HiperSockets Converged Interfaces) settings.

Here is a sample session illustrating how to use:

  $ hsci add enc8410 encb040
  Verifying net dev encb040 and HiperSockets dev enc8410
  Adding hsci8410 with a HiperSockets dev enc8410 and an external \
                                                      dev encb040
  Set enc8410 MAC fe:c2:f4:35:00:12 on encb040 and hsci8410
  Successfully added HSCI interface hsci8410

  $ ip addr add 192.1.2.3/16 dev hsci8410

  $ hsci show
  HSCI      PNET_ID  HiperSockets  External
  ---------------------------------------------
  hsci8410  NET1     enc8410       encb040

For further details, see the changelog.

Webinar: Oracle on IBM Z and LinuxONE

Join us for our webinar on Wednesday, February 24th, 11:00 AM - 12:00 PM EST!

Abstract

The most popular workload running on IBM Z and LinuxONE today is Oracle database. There are many business and technical reasons Clients choose our platform for their mission-critical Oracle workloads. Topics relevant to Oracle that will be covered in the session include: Licensing, Security, Storage, HA and DR.
This session will cover:

  • The high-level value proposition of consolidating Oracle on LinuxONE from Exadata and other x86 platforms, including a discussion of Oracle licensing concepts as they apply to moving Oracle workload to our platform.
  • Technically compare LinuxONE and Exadata, reviewing IBM's approach to running and managing Oracle workload on our platform compared to how Oracle approaches this with Exadata.
  • Migration options/approaches customers may use when moving their Oracle workload(s) from Exadata and other x86 platforms to our platform.

Speakers

  • Terry Ginther, LinuxONE Global Data Serving SME, IBM USA
  • David J Simpson, Certified Oracle and Open Source Database Specialist,  IBM LinuxONE and Z Systems, IBM USA

Registration

Register here. You can check the system requirements here.
After registering, you will receive a confirmation email containing information about joining the webinar.

Replay & Archive

All sessions are recorded. For the archive as well as a replay and handout of this session and all previous webinars see here.

github Organization Change

To better reflect the broader scope of our activities, we have renamed our github organization to become IBM s390 Linux.

Projects currently hosted:

Old links should still work by forwarding accordingly, but we recommend to update links and git remote addresses.

New Release: smc-tools v1.5

smc-tools v1.5 is now available for download here

Highlights in this release:

smcd/smcr: New info Command

Provides a summary of both, the SMC levels supported in the Linux kernel, and also the hardware's capabilities:

  root@tux> smcd info
  Kernel Capabilities
  SMC Version:      2.0
  SMC Hostname:     tux
  SMC-D Features:   v1 v2
  SMC-R Features:   v1

  Hardware Capabilities
  SEID:             IBM-SYSZ-ISMSEID00000000XYZ
  ISM:              v1 v2
  RoCE:             n/a

In this example output, we would be all set for SMC-Dv1 and SMC-Dv2, as we have support in the kernel and in ISM. However, SMC-R would not be available: While we have support in the kernel, we are lacking respective hardware.

New Utility smc_chk

This utility provides a couple of useful tools to check on correct SMC setup:

Verify SMC setup using an arbitrary local or remote service. E.g. start a local service with smc_chk -S and run

  root@tux> smc_chk -C 127.0.0.1 -p 37374
  Test with target IP
127.0.0.1 and port 37374
    Live test (SMC-D and SMC-R)
       Success, using SMC-D

to confirm that the host's SMC setup works, and we can connect using SMC-D as expected.

Or use smc_chk to connect to e.g. a z/OS server's 3270 port to verify SMC setup on both ends:

  root@tux> smc_chk -C 192.168.2.95 -p 23
  Test with target IP
192.168.2.95 and port 23
    Live test (SMC-D and SMC-R)
 
      Failed  (TCP fallback), reasons:

           Client:        0x03010000   Peer does not support SMC

Looks like the peer does not have SMC enabled in this case!

Finally, use smc_chk to display PNET IDs of arbitrary interfaces:

  root@tux> smc_chk -i enc8000
  PNET5


For more information, see the README.

New Release: Linux Kernel 5.10

Linux kernel 5.10 (available here) has been released. Here are the highlights from a Linux on Z perspective:

  • Crypto 
    • Added ability to transform secure private EP11 and CCA EC keys into protected EC private keys.
  • Networking
    • HiperSockets Converged Interface (HSCI): This feature allows to converge a HiperSockets device with a networking adapter (OSA-Express or RoCE Express) into a single logical interface. HiperSockets will be preferred whenever a peer has an HSCI using the same HiperSockets device. Otherwise, the external networking adapter will be used. This way, users can take advantage of the benefits of HiperSockets while retaining external connectivity through a single interface.
    • SMC-Dv2: Version 2 of SMC-D requires an IBM z15 and provides two major improvements:
      1. Removal of the same-subnet restriction: Peers that intend to communicate using SMC-D no longer need to be within the same IP subnet. I.e. any peer that is hosted on the same IBM Z machine can be reached using SMC-Dv2.
      2. No PNET ID configuration required: PNET IDs are no longer required to indicate ISM devices to be used with a certain IP interface. Any ISM device attached to the respective Linux instance with the PNET ID either not set or matching any other network device, and in use by both peers will be used for SMC-Dv2.
  • NVMe Devices 
    • IPL Load Normal Support: If a Linux instance is defined with a large memory size and booted from an NVMe device, time required for a reboot is reduced significantly by omitting memory clearing.
    • Stand-Alone Dump Support: It is now possible to trigger a dump by booting from a respectively prepared NVMe device.
  • Server Time Protocol leap seconds scheduling support, see this blog entry. 

Popular Posts