Asynchronous I/O Parameters The following kernel parameters are used for managing asynchronous I/O operations. The first four are related to POSIX asynchronous I/O operations; the last pertains to open ports between processes and the asynchronous disk-I/O driver: | ''aio_listio_max'' | Specifies how many POSIX asynchronous I/O operations are allowed in a single ''listio()'' call. | | ''aio_max_ops'' | System-wide maximum number of POSIX asynchronous I/O operations that are allowed at any given time. | | ''aio_physmem_pct'' | Maximum total system memory that can be locked for use in POSIX asynchronous I/O operations. | | ''aio_prio_delta_max'' | Maximum priority offset allowed in a POSIX asynchronous I/O control block (''aiocb''). | | ''max_async_ports'' | Maximum number of ports to the asynchronous disk-I/O driver that processes can have open at any given time. |
Configurable Parameters for Kernel-Panic Dumps The following parameters affect dump operations when a kernel panic occurs and parts of system memory are dumped to disk: | ''alwaysdump'' | Specifies which classes of system memory are to be dumped if a kernel panic occurs. | | ''dontdump'' | Specifies which classes of system memory are !!not!! to be dumped if a kernel panic occurs. | | ''initmodmax'' | Maximum number of kernel modules that ''savecrash'' will handle when processing system-crash dump data. | | ''modstrmax'' | Maximum size, in bytes, of the ''savecrash'' kernel-module table which contains module names and their location in the file system. |
Overview of Fiber Channel Kernel Parameters Two kernel configuration parameters pertain to the Fiber Channel SCSI subsystem and communication between the system processor and any peripheral devices that interact with it using Fiber-Channel Protocol (FCP). These parameters are used for adjusting the default amount and type of memory allocated for supporting concurrent FCP read, write, and/or control requests. ''num_tachyon_adapters'' Some of the memory that is allocated for FCP requests must meet certain requirements. The ''num_tachyon_adapters'' parameter is used on systems that do not provide I/O virtual addressing so that memory meeting these requirements can be allocated. The value you specify controls what type of memory is allocated, and specifies the number of Tachyon-based Fiber-Channel adapters in the system. ''max_fcp_reqs'' This parameter sets a limit on the number of concurrent FCP requests that are allowed on an FCP adapter. The default value of 512 requests can be changed by specifying a different value for this parameter. The optimal limit on concurrent requests depends on a number of factors such as configuration, device characteristics, I/O load, host memory, and other values that FCP software cannot easily determine. Configurable File-System Parameters The following parameters control various aspects of file-system management: Buffer Cache Group | ''bufpages'' | Number of 4-Kbyte pages in file-system buffer cache. | | ''dbc_min_pct'' | Mimimum percentage of memory for dynamic buffer cache. | | ''dbc_max_pct'' | Maximum percentage of memory for dynamic buffer cache. | | ''nbuf' | System-wide number of file-system buffer and cache buffer headers. |
Open/Locked Files Group | ''maxfiles'' | Soft limit on how many files a single process can have opened or locked at any given time. | | ''maxfiles_lim'' | Hard limit on how many files a single process can have opened or locked at any given time. | | ''nfile'' | Maximum number of files that can be open simultaneously on the system at any given time. | | ''nflocks'' | Maximum combined total number of file locks that are available system-wide to all processes at any given time. | | ''ninode'' | Maximum number of open inodes that can be in memory. |
Asynchronous Writes Group | ''fs_async'' | Specify synchronous or asynchronous writes of file-system data structures to disk. |
VxFS (Journaled) File-System Parameter | ''vx_ncsize'' | Memory space reserved for VxFS directory path-name cache |
Overview of Mass-Storage Kernel Parameters Configurable parameters related to file system performance fall into the following categories: | File System Buffers | Allocating system physical memory resources for file system buffer cache space. | | Open or Locked Files | Number of files that can be open or locked simultaneously. | | Journal File Systems | Allocate space for the Directory-Name Lookup Cache (DNLC) associated with VxFS file-system inodes. | | Asynchronous Writes | Asynchronous writes to file system allowed or not allowed. |
Configurable Parameters for Logical Volume Manager (LVM) Two configurable kernel parameters are provided that relate to kernel interaction with the Logical Volume Manager: maxvgs Maximum number of volume groups configured by the Logical Volume Manager on the system. no_lvm_disks Flag that notifies the kernel when no logical volumes exist on the system. If set, all file systems coincide with physical disks on the system and physical disk boundaries. The only exception to this is when disks are configured for partitions or part of the disk is reserved for swap and other non-file-system uses. Overview of LVM Operation Logical Volume Manager (LVM) is a subsystem for managing file systems and disk storage space that are structured into logical volumes rather than being restricted to the beginning and end points of a physical disk. Logical volumes can be smaller than the disk or disk array on which they reside, or they can include all or part of several disks or disk arrays. Logical volume boundaries are not required to coincide with the boundaries of physical disks when multiple disks or arrays are used. Managing logical volumes is done by the Logical Volume Manager, not the kernel. However, the kernel contains data structures for each volume group on the system, and the space reserved for LVM data structures must be sufficient to support the number of volume groups that exist on the system. This is done by the ''maxvgs'' kernel configuration parameter. A second parameter, ''no_lvm_disks'', is provided for notifying the kernel at boot time that no logical volumes exist on the system. This saves the system from having to identify and activate logical volumes at boot time. Volume and Group Boundaries Logical volume groups consist of one or more logical volumes. Logical volume boundaries within a volume group can be configured anywhere on a given disk. However, a single disk device cannot be shared between volume groups. Disk arrays configured as RAID (redundant array of independent disks) arrays for data protection are treated as a single disk device and cannot be shared between volume groups. Individual disks in any array that is not configured as a RAID array are treated as individual devices, and individual devices can be assigned to any volume group as desired by the administrator. Configurable Parameters for Memory Paging | ``allocate_fs_swapmap'' | Enable or disable allocation of file-system swap space when ``swapon()'' is called as opposed to allocating swap space when ``malloc()'' is called. Enabling allocation reduces risk of insufficient swap space and is used primarily where high-availabilility is important. | | ``maxswapchunks'' | This parameter, multiplied by ``swchunk'' times the system block size value (``DEV_BSIZE''), determines the combined maximum amount device swap and file-system swap space that can be configured, system-wide. | | ``nswapdev'' | Maximum number of devices, system-wide, that can be used for device swap. Set to match actual system configuration. | | ``nswapfs'' | Maximum number of mounted file systems, system-wide, that can be used for file-system swap. Set to match actual system configuration. | | ``page_text_to_local'' | Enable or disable program text segment swapping to local swap device on cluster client. Increases load time for loading memory with new contents (text is written to local disk instead of discarding and reloading later from server), but reduces network traffic from server to client. | | ``remote_nfs_swap'' | Enable or disable swap to mounted remote NFS file system. Used on cluster clients for swapping to NFS-mounted server file systems. | | ``swapmem_on'' | Enable or disable pseudo-swap allocation. This allows systems with large installed memory to allocate memory space as well as disk swap space for virtual memory use instead of restricting availability to defined disk swap area. | | ``swchunk'' | Defines amount of space allocated for each !!chunk!! of swap area. Chunks are allocated from device to device by the kernel. Changing this parameter requires extensive knowledge of system internals. Without such knowledge, Do not change this parameter from the normal default value. |
Variable-Page-Size Parameters | ``vps_ceiling'' | Defines the maximum system-selected page size if the user does not specify a page size. | | ``vps_chatr_ceiling'' | Defines the maximum page size a user can specify by using the ``chatr'' command in a program. | | ``page_size'' | Defines the default minimum user page-size if no page size is specified using ``chatr''. |
Overview of Memory Paging Parameters Configurable kernel parameters for memory paging enforce operating rules and limits related to virtual memory (swap space). They fall into the following categories: | Total System Swap | Maximum swap space that can be allocated system-wide. | | Device Swap | Swap space allocated on hard disk devices. | | File System Swap | Swap space allocated on mounted file systems. | | Pseudo-Swap | Use of installed RAM as pseudo-swap, allowing virtual memory space allocation beyond the limit of swap space on disk devices. | | Variable Page Sizes | The size of virtual memory pages can be altered to make swap operations more efficient in particular applications. |
Undocumented Kernel Parameters Some of the configurable parameters that appear in the kernel master file are not documented, and some are not known to or are not supported by SAM for any of several possible reasons: * The parameter is obsolete. It is no longer used in current HP-UX releases, but might appear in an existing kernel configuration. If SAM encounters an obsolete parameter in the current kernel configuration, it does not display it in the list of configurable parameters that can be changed. It also removes that parameter when creating the new configuration file used to build the pending kernel. * The parameter is not supported by SAM. As with obsolete parameters, it is not displayed in the list of configurable parameters, but it is retained in the new kernel configuration file to ensure that no malfunctions are introduced due to a missing parameter. * The parameter is assigned a value by kernel configuration software, which is frequently based on external factors. The assigned value might be used when calculating values for one or more other parameters. * The parameter is for HP factory or support use only. No change from the default value should be made unless specifically directed otherwise by official HP support personnel. * The parameter supports obsolete or obsolescent sofware. For information about how to select a non-default value, consult the documentation furnished with the software that the parameter supports.
|