Skip to main content
ONTAP SAN Host Utilities

Configure Oracle Linux 7.0 for FCP and iSCSI with ONTAP storage

Contributors netapp-pcarriga netapp-ranuk netapp-sdaffy netapp-sarajane netapp-reenu netapp-aoife
PDFs

The Linux Host Utilities software provides management and diagnostic tools for Linux hosts that are connected to ONTAP storage. When you install the Linux Host Utilities on an Oracle Linux 7.0 host, you can use the Host Utilities to help you manage FCP and iSCSI protocol operations with ONTAP LUNs.

Note You don't need to manually configure Kernel-based Virtual Machine (KVM) settings because ONTAP LUNs are automatically mapped to the hypervisor.

Step 1: Optionally, enable SAN booting

You can configure your host to use SAN booting to simplify deployment and improve scalability.

Before you begin

Use the Interoperability Matrix Tool to verify that your Linux OS, host bus adapter (HBA), HBA firmware, HBA boot BIOS, and ONTAP version support SAN booting.

Steps

  1. Create a SAN boot LUN and map it to the host.

  2. Enable SAN booting in the server BIOS for the ports to which the SAN boot LUN is mapped.

    For information on how to enable the HBA BIOS, see your vendor-specific documentation.

  3. Verify that the configuration was successful by rebooting the host and verifying that the OS is up and running.

Step 2: Install the Linux Host Utilities

NetApp strongly recommends installing the Linux Host Utilities to support ONTAP LUN management and assist technical support with gathering configuration data.

Install Linux Host Utilities 7.1.

Note Installing the Linux Host Utilities doesn't change any host timeout settings on your Linux host.

Step 3: Confirm the multipath configuration for your host

You can use multipathing with Oracle Linux 7.0 to manage ONTAP LUNs. Oracle Linux 7.0 supports Unbreakable Enterprise Kernel (UEK) R3 and UEK R4. The OS boots with UEK R3 kernel by default

Note You can use the recommended settings for Red Hat Enterprise Linux (RHEL) 7.0 to configure Red Hat Compatible Kernel for Oracle Linux 7.0.
Steps

Depending on your configuration, select the UEK 3 or UEK 4 tab.

UEK 3

The settings for Oracle Linux UEK 3 with and without ALUA update automatically. After the settings update, you need to enable "ALUA Handler":

  1. Create a backup of the initrd-image.

  2. Append the following parameter value to the kernel for ALUA and non-ALUA to work:

    rdloaddriver=scsi_dh_alua
    Example output
    kernel /vmlinuz-3.8.13-68.1.2.el6uek.x86_64 ro root=/dev/mapper/vg_ibmx3550m421096-lv_root rd_NO_LUKSrd_LVM_LV=vg_ibmx3550m421096/lv_root LANG=en_US.UTF-8 rd_NO_MDSYSFONT=latarcyrheb-sun16 crashkernel=256M KEYBOARDTYPE=pc KEYTABLE=us rd_LVM_LV=vg_ibmx3550m421096/lv_swap rd_NO_DM rhgb quiet rdloaddriver=scsi_dh_alua

  3. Recreate the initrd-image:

    dracut -f

  4. Reboot the host.

  5. Review the output of the cat /proc/cmdline command to verify that the setting is complete.

UEK 4

For Oracle Linux UEK 4, verify that the /etc/multipath.conf file is defined and that the NetApp recommended settings are configured for your ONTAP LUNs.

  1. Verify that the /etc/multipath.conf file exits. If the file doesn't exist, create an empty, zero-byte file:

    touch /etc/multipath.conf

  2. The first time the multipath.conf file is created, you might need to enable and start the multipath services to load the recommended settings:

    chkconfig multipathd on
    /etc/init.d/multipathd start

  3. Each time you boot the host, the empty /etc/multipath.conf zero-byte file automatically loads the NetApp recommended host multipath parameters as the default settings. You shouldn't need to make changes to the /etc/multipath.conf file for your host because the operating system is compiled with the multipath parameters that recognize and manage ONTAP LUNs correctly.

    Show parameter settings
    Parameter Setting

    detect_prio

    yes

    dev_loss_tmo

    "infinity"

    failback

    immediate

    fast_io_fail_tmo

    5

    features

    "2 pg_init_retries 50"

    flush_on_last_del

    "yes"

    hardware_handler

    "0"

    no_path_retry

    queue

    path_checker

    "tur"

    path_grouping_policy

    "group_by_prio"

    path_selector

    "service-time 0"

    polling_interval

    5

    prio

    "ontap"

    product

    LUN

    retain_attached_hw_handler

    yes

    rr_weight

    "uniform"

    user_friendly_names

    no

    vendor

    NETAPP

  4. Verify the parameter settings and path status for your ONTAP LUNs:

    multipath -ll

    The default multipath parameters support AFF and FAS configurations. In these configurations, a single ONTAP LUN shouldn't require more than four paths. If there are more than four paths, it might cause issues with the paths during a storage failure.

    An AFF or FAS configuration should have two groups of paths with higher and lower priorities. Higher priority Active/Optimized paths are served by the controller where the aggregate is located. Lower priority paths are active but non-optimized because they are served by a different controller. Non-optimized paths are only used when optimized paths aren’t available.

    The following example outputs show the correct parameter settings and path status for ONTAP LUNs in an AFF or FAS configuration with two Active/Optimized paths and two Active/Non-Optimized paths:

    Show example 
    multipath -ll
3600a0980383036347ffb4d59646c4436 dm-28 NETAPP,LUN C-Mode
size=10G features='3 queue_if_no_path pg_init_retries 50' hwhandler='1 alua' wp=rw
|-+- policy='service-time 0' prio=50 status=active
| |- 16:0:6:35 sdwb  69:624  active ready running
| |- 16:0:5:35 sdun  66:752  active ready running
`-+- policy='service-time 0' prio=10 status=enabled
  |- 15:0:0:35 sdaj  66:48   active ready running
  |- 15:0:1:35 sdbx  68:176  active ready running

Step 4: Optionally, exclude a device from multipathing

If required, you can exclude a device from multipathing by adding the WWID for the unwanted device to the "blacklist" stanza for the multipath.conf file.

Steps

  1. Determine the WWID:

    /lib/udev/scsi_id -gud /dev/sda

    "sda" is the local SCSI disk that you want to add to the blacklist.

    An example WWID is 360030057024d0730239134810c0cb833.

  2. Add the WWID to the "blacklist" stanza:

    blacklist {
	     wwid   360030057024d0730239134810c0cb833
        devnode "^(ram|raw|loop|fd|md|dm-|sr|scd|st)[0-9]*"
        devnode "^hd[a-z]"
        devnode "^cciss.*"
}

Step 5: Customize multipath parameters for ONTAP LUNs

If your host is connected to LUNs from other vendors and any of the multipath parameter settings are overridden, you need to correct them by adding stanzas later in the multipath.conf file that apply specifically to ONTAP LUNs. If you don't do this, the ONTAP LUNs might not work as expected.

Check your /etc/multipath.conf file, especially in the defaults section, for settings that might be overriding the default settings for multipath parameters.

Caution You shouldn't override the recommended parameter settings for ONTAP LUNs. These settings are required for optimal performance of your host configuration. Contact NetApp support, your OS vendor, or both for more information.

The following example shows how to correct an overridden default. In this example, the multipath.conf file defines values for path_checker and no_path_retry that aren't compatible with ONTAP LUNs, and you can't remove these parameters because ONTAP storage arrays are still attached to the host. Instead, you correct the values for path_checker and no_path_retry by adding a device stanza to the multipath.conf file that applies specifically to the ONTAP LUNs.

Show example 
defaults {
   path_checker      readsector0
   no_path_retry     fail
}

devices {
   device {
      vendor          "NETAPP"
      product         "LUN"
      no_path_retry   queue
      path_checker    tur
   }
}

Step 6: Review the known issues

The Oracle Linux 7.0 host with ONTAP storage has the following known issues:

NetApp Bug ID Title Description

901558

OL7.0 : Host loses all paths to the lun and hangs due to "RSCN timeout" error on OL 7.0 UEK r3U5 Beta on Emulex 8G(LPe12002) host

You might observe that the Emulex 8G(LPe12002) host hangs and there is a high I/O outage during storage failover operations with I/O. You might observe paths not recovering, which is a result of the RSCN timeout, due to which the host loses all the paths and hangs. Probability of hitting this issue is high.

901557

OL 7.0: High I/O outage observed on QLogic 8G FC (QLE2562) SAN host during storage failover operations with I/O

You might observe high I/O outage on QLogic 8G FC (QLE2562) host during storage failover operations with I/O. Aborts and Device resets manifests as I/O outage on the host. Probability of hitting this I/O outage is high.

894766

OL7.0: Dracut fails to include scsi_dh_alua.ko module in initramfs on UEKR3U5 alpha

The scsi_dh_alua module might not load even after adding the parameter "rdloaddriver=scsi_dh_alua" in the kernel command line and creating Dracut. As a result, ALUA is not enabled for NetApp LUNs as recommended.

894796

Anaconda displays an iSCSI login failure message although logins are successful during OL 7.0 OS installation

When you are installing OL 7.0, the anaconda installation screen displays that iSCSI login to multiple target IPs have failed though the iSCSI logins are successful. Anaconda displays following error message:
"Node Login Failed"
You will observe this error only when you select multiple target IPs for iSCSI login.
You can continue the OS installation by clicking the "ok" button. This bug does not hamper either the iSCSI or the OL 7.0 OS installation.

894771

OL7.0 : Anaconda does not add bootdev argument in kernel cmd line to set IP address for iSCSI SANboot OS install

Anaconda does not add a bootdev argument in the kernel command line where you set the IPv4 address during the OL 7.0 OS installation on an iSCSI multipath'd LUN. Owing to this, you cannot assign IP addresses to any of the Ethernet interfaces that were configured to establish iSCSI sessions with the storage subsystem during the OL 7.0 boot. Since iSCSI sessions are not established, the root LUN is not discovered when the OS boots and hence the OS boot fails.

916501

Qlogic 10G FCoE (QLE8152) host kernel crash observed during storage failover operations with I/O

You might observe a kernel crash in Qlogic driver module on 10G FCoE Qlogic (QLE8152) host. The crash occurs during storage failover operations with I/O. Probability of hitting this crash is high which leads to longer I/O outage on the host.

What's next?

  • Learn about using the Linux Host Utilities tool.

  • Learn about ASM mirroring.

    Automatic Storage Management (ASM) mirroring might require changes to the Linux multipath settings to allow ASM to recognize a problem and switch over to an alternate failure group. Most ASM configurations on ONTAP use external redundancy, which means that data protection is provided by the external array and ASM doesn't mirror data. Some sites use ASM with normal redundancy to provide two-way mirroring, normally across different sites. See Oracle databases on ONTAP for further information.