Skip to main content
  • Place orders quickly and easily
  • View orders and track your shipping status
  • Enjoy members-only rewards and discounts
  • Create and access a list of your products
  • Manage your Dell EMC sites, products, and product-level contacts using Company Administration.

Dell Solutions Enabler 10.0.1 Release Notes

PDF

Known issues

The following high severity issues remain unresolved in this release.

NOTE:In addition to these release notes, fixes, issues, and limitations (and other Dell product releases) can be viewed using Dell Online Support at https://www.dell.com/support/home.
Table 1. Solutions Enabler V10.0.1 known issues
Issue ID Functional area Description Workaround/resolution
Access Control When using the default hardware-based mechanism for generating access IDs, upgrading or changing host hardware components on some platforms can result in a change to the host access ID. If the host access ID changes, the host may be denied access, or assume a different set of privileges, if the default access ID is present. Dell recommends, as part of our security best practices on all hardware platforms, that the optional alternate access ID mechanism be used instead of the default hardware-based access ID mechanism. For more information about setting up Access Control environments including changing a hosts access ID, refer to Dell Solutions Enabler Installation and Configuration Guide.
AIX-specific When you power down or reboot an AIX host, it will no longer recognize your mapped BCV devices. To fix this problem, you must run a special BCV script (mkbcv). You can run the following command procedure to work around this problem:
 
cd /
./inq.AIX | more (look for no gaps in the numbers, ie.. rhdisk0,
rhdisk1, rhdisk3... - rhdisk2 is missing)
cd /usr/lpp/Symmetrix/bin
./mkbcv -a ALL
cd /
./inq.AIX | more (look for no gaps in the numbers, ie.. rhdisk0,
rhdisk1, rhdisk2... - rhdisk2 is not missing)
It is recommended to have ./mkbcv -a ALL in your AIX boot procedures. The example below show you how to download the above mentioned inquiry and make files from the following Dell ftp locations:
NOTE:Change the directory to the most current release.
Hostname : ftp.emc.com
Username : anonymous

Directory: /pub/symm3000/inquiry/
File : inq.aix...[select file for your system here]

Directory: /pub/symm3000/aix/ODM_DEFINITIONS/
File : mkbcv

If you delete a PowerPath pseudo device (hdiskpower) using rmdevice -d, and then add the native device into a composite group, the device will not persist when you reboot. Also, if you delete a powerpath0 device, all of the non-visible composite group members become lost upon reboot.

When you have to delete these devices, use: rmdevice -l
When performing certain dynamic SRDF operations, the R1/R2 inquiry bit settings might get swapped. Without re-scanning the AIX host to update the ODM configuration, the R1/R2 device definitions are not correct. If these devices are mapped to an AIX host, the ODM database may be left with stale data. Using ODM information, you get the incorrect device attributes and inquiry data. For instance, an R1 device will remain an R1 device node, but it really is an R2 device. SYMAPI software does direct SCSI inquiry to the devices and can still recognize the new device configuration.

The same criteria holds true for configuration change operations where BCV or SRDF device attributes are being changed. The ODM configuration will not reflect the latest changes without device re-scanning on the AIX host.

Even after a rescan on the AIX host, the former R1 devices will appear in the definedstate and new device nodes for the R2 appear in the available state. It is recommended in this case to manually delete the defined devices. On AIX machines, SYMCLI operations may fail because the storapid or storsrvd daemons run out of memory. AIX has a limit on the amount of data that a process can allocate. Therefore, if the daemon is being used heavily, the daemon may run out of memory, even when there is enough physical memory available on the system.

In such cases, the daemons executable can be modified to use a larger amount of memory by issuing the following set of commands:
  1. Stop all Solutions Enabler services: stordaemon shutdown all -immediate
  2. Run the ldedit command (this increases the default data allocation limit of an AIX process from 256 MB to 2 GB). For storapid enter:
    /usr/ccs/bin/ldedit -bmaxdata:0x80000000
    /usr/symcli/daemons/storapid
    For storsrvd enter:
    /usr/ccs/bin/ldedit -bmaxdata:0x80000000
    /usr/symcli/daemons/storsrvd
  3. Restart everything.
NOTE:The command in step 2 can be used for any SYMCLI command needing additional memory.

When experiencing memory issues could be experienced in an AIX 64-bit environment with large objects and/or a large number of extents, you should

Set the AIX-specific environment variable LDR_CNTRL=MAXDATA=VAL to extend the memory limits, where VAL is any value between 0x80000000 and 0xD0000000.
NOTE:Dell EMC does not recommend attempting to resolve a file system larger than 2 TB.
The following is an example of setting the LDR_CNTRL variable to the 0xD0000000@DSA setting, which raises the maximum amount of memory the process can allocate by allowing the process to take what it needs rather than limiting it to a specific value:
LDR_CNTRL=MAXDATA=0xD0000000@DSA
export LDR_CNTRL
<symrslv commands>
unset LDR_CNTRL
NOTE:Dell recommends that you consult with IBM before setting the LDR_CNTRL=MAXDATA=VAL variable.
For VIO Server client hosts, IBM has not yet enhanced the HBA API library to support vSCSI HBAs, therefore syminq hba on a VIO client host does not return HBA information.
Base Management In a HYPERMAX OS 5977 environment, before deleting a device, it must be de-allocated and the de-allocation process must be completed.
The sympd list -vm command on an ESX Server 3.0.2 returns N/A for the VM device name. To correct this, enable the SPC-2 flag. This flag does not need to be enabled on ESX Server 3.5 and later.
The device serial number format is SSNNNNNDDP. The expanded serial number format for HYPERMAX OS 5977 contains no DDP, that is, the DDP value is always 000 (SSNNNNN000). The port number (P) still shows the MSB for devices between 4K and 8K, but can now be ignored because of the expanded serial number format. To determine the correct port number in these cases, subtract 8 from the P field.
<4K device
  • Device Serial Number: 1800014130
  • SymDevName: 014
  • Port: 0
>4K, <8K device
  • Device Serial Number: 1801014138
  • SymDevName: 1014
  • Port: 0
>8K device
  • Device Serial Number: 1802014130
  • SymDevName: 2014
  • Port: 0
When group names are duplicated across a GNS environment (GNS groups of the same name defined on different arrays), SYMAPI attempts to resolve the conflict by renaming the group with an extension to the duplicate group name (such as changing group MyGroup to MyGroup0003). When group names exceed 24 characters, it becomes difficult for SYMAPI to recover. It is recommended that you limit GNS group names to fewer than 20 characters to avoid this issue.
The GNS daemon can be configured using the gns_remote_mirror setting within the daemon_options file, to mirror local device groups to remote Symmetrix arrays. This remote mirroring will not be performed if the remote (via SRDF) Symmetrix is also directly connected to the local host. For example, if local Symmetrix S1 is SRDF-connected to Symmetrix S2 and S2 and is also connected locally to the host, groups on S1 will not be mirrored to S2.
Running the storapid daemon makes it possible to allow non-privileged users (UNIX, non-Root; Windows, non-Administrator users) to execute SYMAPI applications. However, these users must be given proper access rights to the SYMAPI configuration database file.
When VCM is enabled, and all devices are unmasked (not visible) on a port, the symsan command still returns one record. This is because the SCSI specification states that LUN 0 must always respond to SCSI commands on a given port. In this case, the device will be marked Not Ready and may or may not correspond to a mapped device.
Unconfigured (empty) disks do not list properly. When all devices on a disk are removed, the symdisk list command output still shows the disk as not empty. To fix this issue, delete the database (symapi_db.bin) and issue a discover command.
Solutions Enabler may not get an updated state for the blinking door LED for up to 60 seconds. As result the state reported may not reflect real state of the LED. If a change occurs within the last 60 seconds, the changed state may not be reported.
The symchg command cannot be used for remote operations. In addition, the symchg command fails if the following occurs:
  • On IBM AIX platforms, if a physical device is a member of a volume group. In this case, AIX places a SCSI reservation on the device, which prevents symchg from opening the device properly.
  • A failed device is encountered. If symchg delete fails, you must apply the command to one device at a time for all devices of the group.
Configuration Manager A change in the Enginuity reporting of Spindles to Solutions Enabler causes very large IDs to be reported. This causes a significant increase in memory used by Solutions Enabler. The increase is approximately 2 MB for each array.
Device masking An error may return when masking a device if the Symmetrix array director is nonparticipating.
Attempting to restore a backup file may fail to restore all masking data after a configuration change that results in empty groups.

For example, assume you have a port group that contains an FA that was ACLX-enabled when the backup was created. After a configuration change operation, that same FA is no longer ACLX-enabled. This port will be removed from the port group and any view that contained this group will not be restored. if either the SG or the IG within the view belong to another view that is restored, the masking data is not properly restored.

symaccess users can explicitly assign the ACLX LUN to the device that has been provisioned by using the -lun option. In addition, the ACLX LUN can be implicitly assigned to a device if none of the ports in the masking view have the show ACLX port flag enabled when the device is initially provisioned to the host. Additional paths provisioned to the host for that device would then continue to use the same LUN (e.g. the ACLX LUN). This can cause an ACLX device that was visible to the host on those paths to lose its visibility and may cause the provisioning operation to fail if the ACLX device is being used as the gatekeeper for the operation.
OPT599456 If the array has the FILE guest, the restore operation will not be allowed.
Double Checksum When using Veritas VxVM with Double Checksum, rejected I/O may cause plexes to become detached. To return to normal operations, use vxedit to remove the failing status and run vxplex to reattach the plex.
The checksum validate option (-validate) does not collapse the extents reported. This may cause the validate action to incorrectly report that too many extents need to be displayed. Typically this would occur in logical volume environments where more than 31 data files are housed in a single volume group. If the validate action returns an error about the number of extents, it is likely that the enable or disable operations would succeed. Despite this warning, continue with your enable/disable operations.
When discard is used, the I/O is processed in 64 kilobyte chunks. If an I/O is rejected, it is rejected in that granularity as well. Therefore, some good blocks may be rejected along with the bad block(s) if they fall within the same chunk.
An operating system patch is required before using Dell Double Checksum on HP-UX 11.23. The patch addresses a retry looping condition when a bad I/O is rejected by the array. To check and/or fix this problem, consult your HP representative for the Oracle Hard Integrity error patch.
HP-UX-specific For multithreaded environments on HP platforms, you can get a bus error after a pthread_create() call. Increase your host system stack size setting up to 2 MB or more to circumvent these errors.
The syminq hba requires that system log files exist in /var/adm to successfully discover HBAs that are not SNIA compliant. These HBAs may not be reported if the log files are aged, moved, or renamed. Rebooting the host will refresh HBA information in the system logs, allowing syminq to report the HBA information.
Installation and upgrades When using Native installer commands (such as pkgadd, rpm, swinstall, installp, and setld), uninstall previous versions of Solutions Enabler before upgrading to Solutions Enabler V9.2.4.
If the Windows Services Manager application is open while upgrading from a previous version, the installation may fail with the following message: Could not uninstall SYMAPI server. Closing the Services Manager remedies this problem.
When manually creating certificates for secure remote transmissions, there is a possibility that the space in the certificate for server information could be exceeded. Only the hostname can be specified in the Common Name. During the execution of manage_server_cert command you would see the following error: The Common Name may not exceed 55 characters

The space is limited to 64 characters. This includes the reserved token storsrvd that precedes the value specified for the create command. The maximum length of the Common Name argument is 55 characters.

As a workaround, host names or cluster names can be specified in the -san argument. For example, you could specify -san a12345678.internal.abc.com. Refer to the EMC VMAX All Flash and VMAX3 Family Security Configuration Guide for details on modifying the certificate.
When Solutions Enabler is installed, a certificate is created for securing remote transmissions for client/server operations. This certificate is valid for a period of 3650 days. The validity period start date is the current date and time set on the machine on which the certificate is created. If the date and time of the client is significantly behind the server this could cause a handshake error at the client during client/server communications. This problem can be verified on the server by looking at the storsrvd.log in the log directory and seeing if there is a Bad Certificate error. The solution to this problem is to synchronize as closely as possible the date/time of the client and server machines on which storsrvd is running.
Linux on System z-specific When running Linux on System z as a z/VM guest, CKD devices must be defined as unsupported DASD and dedicated to the guest. In addition, the module vmcp must be installed so that Solutions enabler is able to query CP as to how a device is defined. Solutions Enabler will not discover any CKD devices not attached in this manner.
Linux Red Hat/SUSE-specific If the host has 'mapper' devices (/dev/mapper/mpathX), Solutions Enabler requires the libdevmapper.so (64-bit) library to be installed to be able to provide information on these devices.

Usually the library is versioned (libdevmapper.so.1.02), so it must be pointed to by the symbolic link libdevmapper.so.

Refer to Dell Knowledgebase articles 474135 and 438569 for more information.

An intermittent issue within the Linux environment causes a passive system call operation (SCSI 3B/3C sequence where data is returned from the array) to return with a success, but 0 bytes transferred. Because of this issue, Solutions Enabler is throwing different outputs at different time instances for the same command symsnapvx list -sid SymmID -v.
NDM After a symdm commit, the VMAX devices are permanently assigned the VMAX3 and VMAX ALL FLASH device identity (WWN). If the VMAX devices are subsequently deleted and recreated, then the devices will continue to have the VMAX3 and VMAX ALL FLASH device identity, however, the effective device identity will be incorrectly reported by the SYMCLI and Unisphere as having the VMAX device identity. Once the VMAX device is masked to a host, the reporting of the effective device identity will correctly show the VMAX3 and VMAX ALL FLASH device identity.
SRDF The symrdf set nr_if_invalid action does not immediately take effect over the SRDF link for both regular and consistency group environments.
17439068 STP Statistics affected by an ePack are incorrectly reported after the update. Restart the STP daemon to correctly process statistics after the update.
TimeFinder When locally performing a symmir query out to a remote array, the command always returns a 0 value for the invalid tracks of the remote standard device(s) when these devices are in the split state.
When a restore to the R1 device is ready on the link, the -remote option must be specified to acknowledge the remote copy.
When the SYMAPI_WAIT_FOR_BCV_SYNCH option is set to TRUE and large devices are in use, a timeout error can result, during a full or incremental establish, or incremental restore operation.
When operations are performed on a device group, with the -exact option, the pairings of the devices are preserved as they were entered in the device group.

However, when operations are performed on the local devices in a composite group with the -exact option, the exact pairing operation considers the context of the local array when performing the exact pairing of the standard and BCV devices.

When operations are performed on the remote devices in a composite groups with the -exact option, the exact pairing operation considers the context of the array and associated SRDF group when performing the exact pairing of the standard and BCV devices.

For proper TimeFinder/Mirror operations with Veritas (VxVM) environments, you must run: vxdctl enable
For Informix environments, symmir and symrdb list actions might ignore showing physical devices that have empty database tables or indexes.
The Enginuity Consistency Assist (ECA) feature can be used to perform consistent split operations. This feature is implemented using the -consistent option with the symmir split, symclone activate, symclone establish, and symrcopy activate.

The symclone activate command takes advantage of the consistency feature.

To consistently split BCV pairs using ECA you must have either a control host where you are not running data write I/O, or a database host with a dedicated channel to perform the consistent split.

Once the symmir split -consistent command is issued, I/O to the device group is frozen and a 30 second Enginuity protection timer begins and opens an ECA window. After 30 seconds, the I/O channels are thawed, granting (pent up) operations access to the standard devices. Splits across all devices in a group are considered consistent if the BCV split execution is performed within that window. If for some unknown host or I/O channel reason, not all devices are split within the 30 second window, symmir returns the following at completion:

Consistency window was closed on some devices before the operation completed.

At this point, the final successful split outside the window is no longer considered consistent in execution across the device group. For consistency and reliability, you should re-establish the device group and then (later) attempt the consistent split again.

There are possible instances in a client/server environment that the return error message for a failed TimeFinder operation is vague or incomplete, as shown in the following example with no text after the colon: "BCV Library error is:"

When this occurs, view the SYMAPI log file on the server for the complete error message statement.

When using symmir -rrbcv to verify the states with a LDEV list, always specify the standard and target pairs. Only specifying the standard devices in the LDEV list may not return the correct result.

TimeFinder clone copy session in nocopy mode — In HYPERMAX OS 5977 or higher, when activated, the clone nocopy session shares tracks with source and hence will not initiate any copy when reading from target device. Only when writing to the source or target device is the copy initiated.
Virtual Provisioning Following an allocate operation on a device, the total written track counters incorrectly show the entire device as having been written to.
On arrays running HYPERMAX OS 5977 or higher, you can disable devices that have allocated tracks, as long as number of the devices does not exceed twenty percent of the total number of enabled devices (disabled devices are not counted as part of devices in the pool).
z/OS-specific Do not delete the default Access Control group called UnknwGrp. In addition, the access IDs and access groups should not be removed from the UnknwGrp group. If these items are removed, the Symmetrix array will not be properly discovered.
During the SSL handshake process, the SSL-enabled client does a reverse lookup of host name using the IP address of the Solutions Enabler server to which it is connecting. If the fully qualified host name from the lookup does not match the host's fully qualified name, an SSL handshake error will occur. To resolve this, the z/OS certificates may be generated with additional hostnames when the zoscert command is run.

For example, add an alternate hostname like this when prompted by the zoscert command:

Fully qualified z/OS hostname []: a1.host.domain b2.host.domain

The first name specified is used in the Common Name field in the generated certificate. Each succeeding name or IP address is coded in a Subject Alternative Name entry in the certificate.


Rate this content

Accurate
Useful
Easy to understand
Was this article helpful?
0/3000 characters
  Please provide ratings (1-5 stars).
  Please provide ratings (1-5 stars).
  Please provide ratings (1-5 stars).
  Please select whether the article was helpful or not.
  Comments cannot contain these special characters: <>()\