Isilon and PowerScale Upgrade - Customer Preparation Guide v3.7
Isilon and PowerScale Upgrade - Customer Preparation Guide v3.7
In preparation for your upcoming upgrade event, please note the below known items:
• Schedule your upgrade for a time outside of your peak I/O load
• Review the appropriate Upgrade Planning and Process Guide:
o Upgrading from OneFS 7.x
o Upgrading from OneFS 8.x
Special attention should be given to the section titled Client connections during rolling upgrades which outlines
client impacts during activities that reboot nodes in a rolling manner. You can also review Isilon non disruptive
upgrade best practices for more detail info.
• A system job of type Upgrade that upgrades the file system after a software version upgrade, is automatically
started and must be allowed to run until completion after committing a OneFS upgrade. If there are inode structure
changes between the OneFS releases of the upgrade, the Upgrade job performs inode structure changes on every
file. When this occurs, the running Upgrade job may cause temporary performance issues due to the increased IOPS
required to perform inode structure changes. Review KB 194551 for detail information.
• All customer data should be stored in the /ifs partition. Data stored on other partitions on nodes may not be
retained during a OneFS upgrade. If you would like to preserve specific files or directories during an OneFS upgrade,
please refer to KB 093309.
• If you have any custom certificates, they may not persist through a OneFS upgrade, please back them up and be
prepared to reapply them post OneFS upgrade, please also see KB 323665 for more information
• If any custom scripts are running on your Isilon/PowerScale cluster, it is highly recommended to validate them
against the destination OneFS version by utilizing an Isilon OneFS Simulator.
• If a NMDP backup is in use, stop the active NDMP backups 30-60 minutes before the upgrade. Please also ensure to
verify compatibility of your NDMP backup solution with the destination OneFS version by reviewing PowerScale and
Isilon OneFS for Backup & Archive Simple Support Matrix. Contact the backup solution vendor support if further
assistance is needed.
• By default, RPS will commit your code post upgrade per process, and you will not be able to roll back to
previous code version after the commit. If you choose for RPS not to commit post upgrade (8.X only), you
must make it clear to DELL EMC that code is to be left uncommitted. Review KB 304002 for further info.
• If you are upgrading from code prior to OneFS 8.1 to any OneFS 8.1 code or later, following the upgrade, you must
ensure your elicensing is updated and your SRS V3 reconfigured. You can find a quick reference on how to do this on
KB 525065; please ensure to review the full documentation available on the Web Administration Guide relative to
your 8.1.x OneFS and the Licensing and Workflow FAQ.
• If you intend on upgrading from code prior to OneFS 7.2 to any OneFS 7.2 code or later:
o NFSv2 is not compatible with OneFS 7.2 or later.
o If you have NFSv3 clients that require 32-bit cookies, review ETA 205085 and take any appropriate action
prior to the upgrade.
Dell Customer Communication - Confidential
o If you have NFS clients, review ETA 483840 and complete steps in the resolution to determine if this may put
access to your data at risk
• In OneFS versions 8.0 - 8.2.1, NFSv4 clients may experience periods of data unavailability during rolling upgrades
due to an issue with Persistent Registers. For more details on this issue and workarounds to prevent the
disruptions please review KB 532616.
• SyncIQ replications must be halted before upgrading the OneFS operating system.
• If you're upgrading your cluster from a version prior to OneFS 8.2.x, and your cluster is in Compliance mode, you
must upgrade both the SyncIQ source and SyncIQ target clusters to the same version of OneFS and patches before
restarting SyncIQ backups. If the SyncIQ source and SyncIQ target clusters are not on the same versions, SyncIQ
backups will fail. This issue is resolved in OneFS 8.2.x
• The SmartConnect Service IP (SSIP) was never intended for client connectivity or WebUI/CLI administration. In
OneFS 8.2 and later, this is strictly enforced and the SSIP is not usable for client or administrative CLI/WebUI access
to the cluster. For more details please review KB 536650
• For information about OneFS compatibility with Hadoop, review the Hadoop - Isilon Info Hub page
• If you are planning an upgrade to from pre-OneFS 8.2 to OneFS 8.2 and later with CloudPools in use, please
make sure to review KB 544023 for the additional steps and information for the upgrade.
• If your OneFS 9.0.0.0 cluster includes PowerScale F200 or PowerScale F600 nodes, and you have SRS v2
installed, the SRS service might not work properly. Please upgrade to SRS v3 before enabling the SRS service on
clusters that include the PowerScale F200 or PowerScale F600 nodes.
• If you are using connectemc service with "isi remotesupport connectemc" enabled, this method must be disabled
before upgrading to OneFS 9.1 or later. It is recommended to configure the SRSv3 method for remote connectivity
via "isi esrs" commands or in the WebUI. See How to Configure an Isilon Cluster to
• Communicate with SRS v3 Gateway Servers for more information on how to configure SRSv3.
• Make sure "Gratuitous ARP Flooding" is enabled on external switches, or clients may experience intermittent
disconnects when SmartConnect moves IPs around during reboots. See KB 531354 for detail information.
• The total amount of available space on the entire cluster must be at least 10 percent, and the total free space for
each node must be at least 8 percent. Do not continue with the upgrade if the minimum available-space
requirements are not met. For more information, see Maintaining Enough Free Space on Isilon Clusters and Pools.
• The latest information relating to Isilon/PowerScale upgrade issues is available at KB 303954.
Additional Information:
• For awareness: It generally takes several business days for all scheduling and pre-work to be completed. However,
we will do our best to make this process as straightforward as possible. The upgrade lead time (average wait for an
available upgrade slot) for each of our supported products can be viewed here.
• If you have multi upgrade requests, please make our scheduling team aware so that all activities can be coordinated.
• If SRS Gateway is in place, please make sure to review KB 525338 as well prior to any code upgrade.
• Dell EMC Isilon HealthCheck is a smart, fast and effective application that enables you to self-support common Isilon
issues and accelerate time-to-resolution. This is the same application used by EMC Isilon Technical Support
Engineers and Field Representatives to resolve service requests. Use it to diagnose, troubleshoot and proactively
avoid issues by analyzing the current health of your cluster and listing items that require attention.
• In advance of performing an upgrade event It is strongly suggested that a business continuance option is available in
the unlikely event it is needed. Dell Technologies Services recommends strongly to implement such a solution prior to
performing any code upgrade.
• You will be requested during initial scheduling to confirm your backup/restore capabilities