Custom Reports Using Data Collectors
Product: Storage Foundation Cluster File System
Platform: AIX
Product version: 5.1
Product component: File System
Check category: All
Check category: Availability
Check description: Checks for the use of instant snapshots and SmartMove. Data corruption can occur with Storage Foundation 5.1 when you use instant snapshots and SmartMove with certain combinations of I/O and administrative operations.
Check procedure:
Check recommendation: If you plan to use instant snapshots (either full-sized or space-optimized), turn off SmartMove. Enter:
# vxdefault set usefssmartmove none
Do one of the following:
# vxdefault list usefssmartmove
Learn More...
About SmartMoveCheck category: Availability
Check description: Checks for the presence of VCS replication agents on the system.
Check procedure:
Check recommendation: It is recommended that you install the missing VCS replication agents listed in the output details.
Check category: Availability
Check description: Checks whether a valid fsck policy has been specified for all the Mount resources that are in the offline state to automatically recover the file systems.
Check procedure:
Check recommendation: Set the FsckOpt attribute for the affected Mount resource to either -Y (fix errors during fsck) or -N (do not fix errors during fsck).
Check category: Availability
Check description: Checks whether the specified mount point is available for mounting after failover happens.
Check procedure:
Check recommendation: If the mount point is mounted, unmount it. Enter the following command: # umount mount_point.
Check category: Availability
Check description: Verifies that the available mount point is not configured to mount a file when the system starts.
Check procedure:
Check recommendation: On a cluster node, make sure that the operating system-specific file system table file does not contain an entry for the mount point. These files are /etc/filesystems (AIX), /etc/fstab (HP-UX and Linux), and /etc/vfstab (Solaris).
Check category: Availability
Check description: Checks whether the specified mount point existing on a cluster node is available for mounting.
Check procedure:
Check recommendation: Create the specified mount point, and make sure that is it not in use.
Check category: Availability
Check description: Checks whether the File System (VxFS) installed on the cluster system where the Mount resource is currently offline has a valid license.
Check procedure:
Check recommendation: Use the /opt/VRTS/bin/vxlicinst utility to install a valid VxFS license on the target cluster systems.
Check category: Availability
Check description: Check whether the File System(VxFS)s mounted on a system are marked for a full file system check (fsck)
Check procedure:
Check recommendation: Repairing file systems that require a full fsck is a time-consuming operation; the time required is proportional to the size of the file system. These files systems will require a full fsck before they can be mounted again; plan any downtime accordingly.
Learn More...
fsck: manual pageCheck category: Availability
Check description: Checks whether enough memory is available on the system for a full file system check (fsck) to run on a mounted File System(VxFS).
Check procedure:
Check recommendation: You do not have enough physical and virtual memory to run a full file system check (fsck) of this file system. In most situations, VxFS replays the intent log, avoiding the need for a full fsck. In rare circumstances, however, a full fsck is required. If you are not at or above Storage Foundation version 5.0 MP3, consider upgrading. 5.0MP3 and higher have reduced memory requirements for a full fsck. Alternately, you can add physical or virtual memory. Physical memory is preferred because using swap space increases the time to complete the check.
Learn More...
fsck: manual pageCheck category: Availability
Check description: Checks whether all the Storage Checkpoints of the mounted VxFS File Systems are removable.
Check procedure:
Check recommendation: The Storage Checkpoints are not removable. In most configurations, Storage Checkpoints should be removable to reduce the chance of an ENOSPC error on the primary data. To create a removable Storage Checkpoint, enter :
#fsckptadm -r create <checkpoint> <mountpoint>
Check category: Availability
Check description: Checks whether the File System disk layout version is supported with the installed Storage Foundation / InfoScale version, and whether the file system size is close to the maximum supported by the disk layout version and block size.
Check procedure:
Check recommendation: The recommendations are summarized in the following two cases:
i) Case 1 : Ensure that disk layout version of any VxFS File Systems mounted on the system are supported by the installed Storage Foundation / InfoScale software. Once you upgrade a disk layout, you cannot downgrade it. Refer to the upgrade recommendations below.
ii) Case 2 : Ensure that the VxFS File Systems mounted on the system are not approaching the maximum size allowed by that disk layout version and block size. It is recommended upgrading the file system disk layout version. Refer to the following table to view the maximum file system size supported by various file system disk layout versions:
FS block size (in K) | Maximum FS size supported (in TB) for FS disk layout Version 5 | Maximum FS size supported (in TB) for FS disk layout Version 6 and higher |
---|---|---|
1 | 4 | 32 |
2 | 8 | 64 |
4 | 16 | 128 |
8 | 32 | 256 |
Learn More...
About disk layoutsCheck category: Availability
Check description: Checks for unmounted VxFS File System(s) that have entries in filesystem table and valid underlying devices. Note: Following platform specific filesystem table are referred in the check: Linux: /etc/fstab AIX: /etc/filesystems HP-UX: /etc/fstab Solaris: /etc/vfstab
Check procedure:
Check recommendation: It is recommended that removing any stale entries and deleting the corresponding underlying volumes to reclaim space.
Learn More...
Editing the vfstab fileCheck category: Availability
Check description: Checks that packages installed across all the nodes in a cluster are consistent
Check procedure:
Check recommendation: Ensure that packages installed on all the nodes in a cluster are consistent and package versions are identical. Inconsistent packages can cause errors in application fail-over.
Check category: Best practices
Check description: Checks that the buffered I/O using I/O sizes doesn't exceed 128K for SF 5.1 and 5.1SP1 on AIX platform. It has recently been determined that VxFS buffered I/O on the AIX platform running Storage Foundation release 5.1 and 5.1SP1 can generate a system panic if the I/O size submitted to the virtual memory management layer is greater than 128Kb.
Check procedure:
Check recommendation: It is recommended that you set the VxFS buffered I/O size to less than 128 KB or upgrade to VxFS version 5.1SP1 and higher which resolves this issue. Please refer to the technote in documentation below for more details.
Check category: Best practices
Check description: Checks VxFS File System fragmentation. If a file system is too fragmented, the check recommends defragmentation.
Check procedure:
Check recommendation: You should defragment the file systems to improve performance and reduce recovery time.
Defragmentation creates I/O and uses CPU, so you should schedule defragmentation during periods of low system activity. A conservative approach is to defragment one file system at a time, that is, to wait for one defragmentation to complete before starting the next. Defragmentation time varies depending on factors including the file system size and the number of files. Defragmentation can run for extended periods.
Learn More...
fsadm_vxfs: manual pageCheck category: Best practices
Check description: Checks whether the installed Storage Foundation / InfoScale products are at the latest software patch level.
Check procedure:
Check recommendation: To avoid known risks or issues, it is recommended that you install the latest versions of the Storage Foundation / InfoScale products on the system.
Check category: Performance
Check description: Checks whether Solid State Drives (SSDs) or flash drives are attached to the server. It also recommends the right version of Storage Foundation and High Availability / InfoScale software that have the SmartIO feature to bring better performance, reduced storage cost and better storage utilization.
Check procedure:
Check recommendation: The recommendation is summarized in below cases:
Case 1 : SSDs or flash drives are detected on the Linux system with the Storage Foundation software version earlier than 6.1 installed. It is recommended to upgrade the Storage Foundation software to 6.1 or higher version which enables you to use the SmartIO feature, which improves performance, reduces storage costs and brings better storage utilization for the applications running on the servers.
Case 2 : SSDs or flash drives are detected on the AIX/Solaris system with the Storage Foundation software version earlier than 6.2 installed. It is recommended to upgrade the Storage Foundation software to 6.2 or higher version which enables you to use the SmartIO feature, which improves performance, reduces storage costs, and brings better storage utilization for the applications running on the servers.
Case 3 : SSDs or flash drives are detected on the Linux system with Storage Foundation software version 6.1 installed, but SmartIO feature is not detected. It is recommended that you use the SmartIO feature which improves performance, reduces storage costs, and brings better storage utilization for the applications running on the servers. Please refer the documentation link(s).
Case 4 : SSD or flash drives are detected on the system with Storage Foundation software version 6.2 or higher installed, but SmartIO feature is not detected. It is recommended that you use the SmartIO feature which improves performance, reduces storage costs, and brings better storage utilization for the applications running on the servers. Please refer the documentation link(s).
Case 5 : Storage Foundation software version 6.2 or higher is found on the AIX/Linux/Solaris system without any SSD or flash drives. SSDs or flash drives are more efficient since they provide faster data access and have a smaller footprint than traditional spinning disks. The data center uses solid-state technologies in many form factors: in-server, all flash arrays, all flash appliances, and mixed with traditional HDD arrays. Each form factor offers a different value proposition. SSDs also have many connectivity types: PCIe, FC, SATA, and SAS. It is recommended that you use the SmartIO feature that offers data efficiency on your SSDs through I/O caching, which improves performance, reduces storage costs, and brings better storage utilization for the applications running on the servers.
Case 6 : Storage Foundation software version 6.1 is found on the Linux system without any SSDs or flash drives. SSDs or flash drives are more efficient since they provide faster data access and have a smaller footprint than traditional spinning disks. The data center uses solid-state technologies in many form factors: in-server, all flash arrays, all flash appliances, and mixed with traditional HDD arrays. Each form factor offers a different value proposition. SSDs also have many connectivity types: PCIe, FC, SATA, and SAS. It is recommended that you use the SmartIO feature that offers data efficiency on your SSDs through I/O caching, which improves performance, reduces storage costs, and brings better storage utilization for the applications running on the servers.
Case 7 : Storage Foundation software version 6.1 is found on the AIX/Solaris system without any SSDs or flash drives. It is recommended to upgrade the Storage Foundation software to 6.2 or higher version and use the SmartIO feature that offers data efficiency on your SSDs through I/O caching which improves performance, reduces storage costs, and brings better storage utilization for the applications running on the servers. SSDs or flash drives are more efficient since they provide faster data access and have a smaller footprint than traditional spinning disks. The data center uses solid-state technologies in many form factors: in-server, all flash arrays, all flash appliances, and mixed with traditional HDD arrays. Each form factor offers a different value proposition. SSDs also have many connectivity types: PCIe, FC, SATA, and SAS.
Check category: Performance
Check description: Checks whether or not the Oracle database is using Veritas Extension for Oracle Disk Manager (ODM) for the VxFS File System. ODM is an API jointly developed by Oracle and Veritas. ODM improves Oracle database performance on file systems.
Check procedure:
Check recommendation: You can use the ODM Extension for VxFS File System to provide better I/O performance to applications using Oracle9i or later. This feature is recommended over the Quick IO (QIO) or Concurrent IO (CIO) features for file system. As of now, this check does not check for QIO/CIO usage.
Check category: Performance
Check description: If the disk layout is version 6 or later, this check compares the size of the VxFS File System and the size of the intent log. If the intent log is too small compared to the VxFS file system, the report recommends resizing the intent log to meet the standards. File System (VxFS) uses an intent log to improve recoverability while maintaining performance. When VxFS creates a file system, it chooses the intent log size. The larger the VxFS file system, the larger the intent log. If you resize the VxFS file system, the intent log size does not change, This can result in a log size smaller than the recommended default. Metadata-intensive workloads can benefit from a larger intent log. This is because if a high number of transactions come in a short period of time, the log fills and must be flushed before accepting new transactions. Metadata-intensive workloads usually add, delete, append to, or truncate files; change file names, permissions, directories, access control lists (ACLs), or owners; or use Storage Checkpoints. Database workloads which pre-allocate large files and only read and write to those files are not usually metadata-intensive -- unless you use Storage Checkpoints. Therefore, they are less likely to see performance benefits from a larger log. You can resize a VxFS file system when it is mounted, online, and actively receiving application I/O with the vxresize command.
Check procedure:
Check recommendation: The VxFS file system(s) have undersized intent logs, which can impact performance. It is recommended to increase your intent log size to meet the standards.
Learn More...
Intent log sizeCheck category: Utilization
Check description: Determines whether the storage enclosure on your system is ready to use the Storage Foundation / InfoScale thin provisioning feature.
Check procedure:
Check recommendation: The recommendations are:
Case I: The VxFS file system resides on thin provisioned storage, but the system does not have a Storage Foundation Enterprise / InfoScale Storage license installed. You need a Storage Foundation Enterprise / InfoScale Storage license to mirror the storage.
Case II: The VxFS file system storage enclosure appears to support thin provisioning but the Storage Foundation / InfoScale software does not detect it as thin provisioned storage. Ensure you have the correct Array Support Libraries (ASLs) installed for this storage.
Case III: Your system appears to be attached to a storage enclosure that supports thin provisioning, and the necessary Storage Foundation / InfoScale product is installed; however, the VxFS file System does not reside on thin provisioned storage. Possible reasons are:
a) Thin provisioning is not used.
b) Thin provisioning is not enabled on the storage enclosure.
c) Your version of the storage enclosure may be an old version that does not support thin provisioning.
For reasons (a) or (b), check the thin provisioning support with your storage vendor.
For reason (c), if you are considering migrating to thin provisioned storage, consider the following:
Note: If you have 5.0 MP3 with HF1, you may not need RP1.
Case IV: The VxFS file system's disk group version is less than 110. It is recommended that you upgrade to a disk group version greater than 110 before attempting to migrate.
Case V: The Storage Foundation / InfoScale with thin provisioning feature does not support the storage enclosure on which the VxFS file system resides.
Learn More...
White Paper on: Storage Foundation / InfoScale and Thin ProvisioningCheck category: Utilization
Check description: Checks for VxFS File System Storage Checkpoints that are older than !param!HC_CHK_FS_OLD_CHECKPOINT_DAYS_OLD!/param!, which is set in the sortdc.conf file.
Check procedure:
Check recommendation: It is recommended that you delete any old VxFS File System Storage Checkpoints that you no longer require in order to reclaim file system space.
Check category: Utilization
Check description: Checks VxFS File System utilization. It lists out the VxFS File Systems whose percentage of usage space is less or more than the percentage specified in the user-defined parameter !param!HC_CHK_FS_USAGE_PERCENT_MIN!/param! and !param!HC_CHK_FS_USAGE_PERCENT_MAX!/param! respectively. These two parameters are set in sortdc.conf file.
Check procedure:
Check recommendation: The VxFS file system(s) listed in output are either under-utilized or over-utilized. It is recommended that you shrink the under-utilized VxFS file system and its volume, and use the freed space elsewhere. It is better to defragment the over-utilized VxFS file systems and to add extra storage.
Check category: Utilization
Check description: Checks if any of the VxFS file system and underlying VxFS volume size is different and the size difference is greater than the size specified in the user-defined parameter !param!HC_CHK_FS_VOLS_SIZE_DIFF_THRESHOLD!/param!, which is set in the sortdc.conf file.
Check procedure:
Check recommendation: To make the best use of volume space, the file system should be the same size as the volume or volume set.
The failure can be summarized in one of the following cases:
Case I: The file system size is less than underlying volume by a threshold parameter of HC_CHK_FS_VOLS_SIZE_DIFF_THRESHOLD.
You should either grow the file system using the fsadm command or shrink the volume using the vxassist command.
Case II: The file system is larger than underlying volume. This can happen due to execution of incorrect command (vxassist) for shrinking the volume after file system creation.
Run the following commands.
To grow the file system:
# fsadm [-F vxfs] [-b <newsize>] [-r rawdev] mount_point
To shrink :
#vxassist -g <mydg> shrinkby <vol> <len>
or
#vxassist -g <mydg> shrinkto <vol> <newlen>
Check category: Utilization
Check description: For multi-volume file systems with storage tiering, checks whether any tier is full or low on space.
Check procedure:
Check recommendation: The storage tier for the multi-volume file system has little available space on it. It is recommended adding more volumes to this tier using the vxvoladm command.
Learn More...
About Multi Volume Filesystem