diff options
Diffstat (limited to 'src/pmdas/solaris/help')
| -rw-r--r-- | src/pmdas/solaris/help | 729 |
1 files changed, 729 insertions, 0 deletions
diff --git a/src/pmdas/solaris/help b/src/pmdas/solaris/help new file mode 100644 index 0000000..4940247 --- /dev/null +++ b/src/pmdas/solaris/help @@ -0,0 +1,729 @@ +# +# Copyright (c) 2000-2004 Silicon Graphics, Inc. All Rights Reserved. +# +# This program is free software; you can redistribute it and/or modify it +# under the terms of the GNU General Public License as published by the +# Free Software Foundation; either version 2 of the License, or (at your +# option) any later version. +# +# This program is distributed in the hope that it will be useful, but +# WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY +# or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License +# for more details. +# +# You should have received a copy of the GNU General Public License along +# with this program; if not, write to the Free Software Foundation, Inc., +# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA +# +# Solaris PMDA help file in the ASCII format +# +# lines beginning with a # are ignored +# lines beginning @ introduce a new entry of the form +# @ metric_name oneline-text +# help test goes +# here over multiple lines +# ... +# +# the metric_name is decoded against the default PMNS -- as a special case, +# a name of the form NNN.MM (for numeric NNN and MM) is interpreted as an +# instance domain identification, and the text describes the instance domain +# +# blank lines before the @ line are ignored +# + +@ kernel.all.cpu.idle Amount of time CPUs were idle +@ kernel.all.cpu.user Amount of time spent executing userspace tasks +@ kernel.all.cpu.sys Amount of time spent executing kernel code +@ kernel.all.cpu.wait.total Amount of time CPU spent waiting for events +@ kernel.percpu.cpu.user Amount of time spent executing userspace tasks by each CPU +@ kernel.percpu.cpu.idle Amount of time each CPU was idle +@ kernel.percpu.cpu.sys Amount of time each CPU spent executing kernel code +@ kernel.percpu.cpu.wait.total Amount of time each CPU spent waiting for events +@ disk.all.read Number of read requests aggregated across all disks +@ disk.all.write Number of write requests aggregated across all disks +@ disk.all.total Number of IO requests aggregated across all disks +@ disk.all.read_bytes Number of bytes read from all disks +@ disk.all.write_bytes Number of bytes written to all disk +@ disk.all.total_bytes Number of bytes transferred to and from all disks +@ disk.dev.read Number of read requests for each individual disk +@ disk.dev.write Number of write requests for each individual disk +@ disk.dev.total Number of IO requests for each individual disk +@ disk.dev.read_bytes Number of bytes read from each individual disk +@ disk.dev.write_bytes Number of bytes written to each individual disk +@ disk.dev.total_bytes Number of bytes transferred to and from each individual disk +@ network.interface.mtu Maximum Transmission Unit of a network interface +Maximum Transmision Unit is the largest size of IP datagram which can be +transferred over the data link. +@ network.interface.in.bytes Number of bytes received by a network interface +@ network.interface.in.errors Number of receive errors per network interface +Number of receive errors per network interface. The errors counted towards +this metric are: IP header errors, packets larger then the link MTU, packets +delivered to the unknown address, packets sent to the unknown IP protocol, +truncated packets, packets discarded due to not having a route for +the destination. +@ network.interface.in.drops Number of packets droped by a network interface +Number of packets discared due to lack of space during the input processing. +@ network.interface.in.delivers Number of packets delivered to ULPs +Number of packets delivered for further processing by the upper-layer +protocols. +@ network.interface.in.bcasts Number of broadcast packets received by a network interface +@ network.interface.in.packets Number of IP packets received by a network interface +@ network.interface.in.mcasts Number of multicast packets received by a network interface +@ network.interface.out.packets Number of packets sent by a network interface +@ network.interface.out.bytes Number of bytes sent by a network interface +@ network.interface.out.errors Number of send errors per network interface +@ network.interface.out.bcasts Number of broadcast packets sent by a network interface +@ network.interface.out.mcasts Number of multicast packets sent by a network interface +@ network.interface.out.drops Number of packets discared by a network interface +Number of packets discared due to lack of space during the output processing. +@ network.udp.ipackets Number of UDP packets received +@ network.udp.opackets Number of UDP packets sent +@ network.udp.ierrors Number of receive errors in UDP processing +@ network.udp.oerrors Number of send erros in UDP processing +@ network.udp.noports Number of UDP packets received on unknown UDP port +Number of UDP packets received for which here is no port can be found. +This counter is reported on the per-interface basis and aggregated by the PMDA. + +@ network.udp.overflows Number of UDP packets droped due to queue overflow +Number of UDP packets droped due to queue overflow. +This counter is reported on the per-interface basis and aggregated by the PMDA. + +@zpool.capacity Total capacity of a zpool in bytes +@zpool.used Total space used on a pool +@zpool.checksum_errors Number of checksum errors per zpool +@zpool.self_healed Number of bytes healed +@zpool.in.bytes Counter of bytes read from a zpool +@zpool.out.bytes Counter of bytes written to a zpool +@zpool.in.ops Counter of reads per zpool +@zpool.out.ops Counter of writes per zpool +@zpool.in.errors Counter of read errors per zpool +@zpool.out.errors Counter of write errors per zpool +@zpool.state Current state of zpool +@zpool.state_int vs_aux << 8 | vs_state + +@zfs.available Amount of space available to the dataset +The amount of space available to the dataset (a filesystem, +a snapshot or a volume) and all its children. This is usually +the amount of space available in the zpool which houses the +dataset. + +@zfs.used.total Amount of space used by the dataset and its children +The amount of space consumed by the filesystem, snapshot or +volume and all its children. This amount does not include +any reservation made by the dataset itself but do include +the reservation of the children. + +@zfs.used.byme Amount of space used by the dataset itself. +This amount exclude any space used by the children of this dataset +or any of its snapshots. + +@zfs.used.bysnapshots Amount of space used by the snapshots of the dataset +The amount of space consumed by the snapshots of this dataset. + +@zfs.used.bychildren Amount of space used by decendents of the dataset +The amount of space consumed by all the decendants of this dataset. + +@zfs.quota Maximum amount of space a dataset can use +Quotas are used to restrict the growth of the datasets. If +the quota is set to 0 then the size of the dataset is limited only +by the size of the pool which houses this dataset. + +@zfs.reservation Minimum amount of space guaranteed to a dataset +The amount of space which dataset and its decendents are guaranteed +to be available for them to use. This amount of taken off the quota +of the parent of the dataset. + +@zfs.compression Compression ratio of the dataset +Compression ratio is expressed as multiplier. To estimate how much data +will be used by the uncompressed data multiply the amount of space used +by the dataset by the compression ratio. + +@zfs.copies Number of redundant copies of data +The number of redundant copies does not include any copies made as +part of the pool redundancy. + +@zfs.recordsize Recommendend block size for files in filesystems +By using recommended block size applications which deal with fixed size +records can improve I/O performance. + +@zfs.used.byrefreservation Space used by refreservation +The amount of space used by a refreservation set on this +filesystem, which would be freed if the refreservation was +removed. + +@zfs.refreservation Minimum amount of space guaranteed to a filesystem +The minimum amount of space guaranteed to a dataset, not +including its descendents. Unlike reservation refreservation is +counted towards the total used space of a dataset. + +@zfs.refquota Amount of space a filesystem can consume +The hard limit on the amount of space a filesystem but not its descendants +can consume from the pool. + +@zfs.referenced Amount of space referenced by the filesystem +The amount of data that is accessible by the filesystem. The data +may be shared with other datasets in the pool. + +@zfs.nsnapshots Number of snapshots in the filesystem + +@zfs.snapshot.compression Compression ratio of the data in the snapshot +Compression ratio is expressed as multiplier. To estimate how much data +will be used by the uncompressed data multiply the amount of space used +by the snapshot by the compression ratio. + +@zfs.snapshot.used Amount of space used by the snapshot + +@zfs.snapshot.referenced Amount of space referenced by the snapshot +The amount of data that is accessible by the snapshot. The data +may be shared with other datasets in the filesystem. + + +@zpool.perdisk.state Current state per disk in zpool +@zpool.perdisk.state_int vs_aux << 8 | vs_state +@zpool.perdisk.checksum_errors Number of checksum errors per disk in zpool +@zpool.perdisk.self_healed Number of bytes healed per disk in zpool +@zpool.perdisk.in.errors Counter of read errors per disk in zpool +@zpool.perdisk.out.errors Counter of write errors per disk in zpool + +@network.link.in.errors Number of input errors per link +Counts input errors per link +@network.link.in.packets Numbe of datagrams received by a link +@network.link.in.bytes Number of bytes received by a link +Counts number of bytes received by a link. For the physical links +this is the raw counter of bytes received, for the aggregated links +this is the number of bytes received by all links in the aggregation +group +@network.link.in.bcasts Number of broadcast datagrams received by a link +@network.link.in.mcasts Number of multicast datagrams +Counts multicast datagram recieved by a link. +@network.link.in.nobufs Number of inpit packets discared +Counts number of packets discared because of failure to allocate buffers +@network.link.out.errors Number of output errors per link +@network.link.out.packets Number of packets sent from a link +@network.link.out.bytes Number of bytes sent from a link +@network.link.out.bcasts Number of broadcast datagrams sent from a link +@network.link.out.mcasts Number of multicast datagrams sent from a link +@network.link.out.nobufs Number of output packets discared +Counts number of packets discared because of failure to allocate buffers +@network.link.collisions Number of collisions detected per link +@network.link.state Link state +1 - Link is up, 2 - Link is down, 0 - unknown state +@network.link.duplex Link duplex +1 - Half duplex, 2 - Full duplex +@network.link.speed Link speed in bytes per second +@hinv.pagesize Memory page size +The memory page size of the running kernel in bytes. +@hinv.physmem Total physical system memory +Total physical system memory size rounded down to the nearest page size +boundary +@pmda.uname identity and type of current system +Identity and type of current system. The concatenation of the values +returned from utsname(2), also similar to uname -a. +@kernel.fsflush.scanned Number of pages scanned by fsflush daemon +@kernel.fsflush.examined Number of pages examined by fsflush daemon +@kernel.fsflush.coalesced Number of pages coalesced into larger page +@kernel.fsflush.modified Number of modified pages written to disk +@kernel.fsflush.locked Number of pages locked by fsflush daemon +Pages which were considered to be on interest for further examination +are locked before deciding if they could be coalesced, released or flushed +to disk. +@kernel.fsflush.released Number of free pages released by fsflush daemon +@kernel.fsflush.time Amount of time fsflush daemon spent doing its work +@mem.physmem Total physical system memory +Total physical system memory size rounded down to the nearest page size +boundary. This metric is the same as hinv.physmem but uses different +units. +@mem.freemem Amount of free memory in the system +@mem.lotsfree Paging theshold +If freemem fails below the lostfree threshold then paging out daemon +starts its activity. Default value for lotsfree is 1/64 of physical memory +or 512K (which ever is larger). +@mem.availrmem Amount of resident memory in the system + +@kernel.all.io.bread Physical block reads across all CPUs +This metric is only updated if reading or writing to UFS mounted filesystems, +reads and writes to ZFS do not update this metric. +@kernel.all.io.bwrite Physical block writes across all CPUs +This metric is only updated if reading or writing to UFS mounted filesystems, +reads and writes to ZFS do not update this metric. +@kernel.all.io.lread Logical block reads across all CPUs +This metric is only updated if reading or writing to UFS mounted filesystems, +reads and writes to ZFS do not update this metric. +@kernel.all.io.lwrite Logical block writes across all CPUs +This metric is only updated if reading or writing to UFS mounted filesystems, +reads and writes to ZFS do not update this metric. +@kernel.all.io.phread Raw I/O reads across all CPUs +@kernel.all.io.phwrite Raw I/O writes across all CPUs +@kernel.all.io.intr Device interrupts across all CPUs + +@kernel.percpu.io.bread Physical block reads +This metric is only updated if reading or writing to UFS mounted filesystems, +reads and writes to ZFS do not update this metric. +@kernel.percpu.io.bwrite Physical block writes +This metric is only updated if reading or writing to UFS mounted filesystems, +reads and writes to ZFS do not update this metric. +@kernel.percpu.io.lread Logical block reads +This metric is only updated if reading or writing to UFS mounted filesystems, +reads and writes to ZFS do not update this metric. +@kernel.percpu.io.lwrite Logical block writes +This metric is only updated if reading or writing to UFS mounted filesystems, +reads and writes to ZFS do not update this metric. +@kernel.percpu.io.phread Raw I/O reads +@kernel.percpu.io.phwrite Raw I/O writes +@kernel.percpu.io.intr Device interrupts + +@hinv.ncpu Number of CPUs in the system +@hinv.ndisk Number of disks in the system + +@kernel.all.trap Traps across all CPUs +@kernel.all.pswitch Context switches across all CPUs +@kernel.all.syscall Total number of system calls across all CPUs +@kernel.all.sysexec Total number of calls from exec(2) family across all CPUs +@kernel.all.sysfork Total number of new processes created across all CPUs +@kernel.all.sysvfork Total number of new processes created across all CPUs +Unlike fork vfork does not copy all the virtual memory of the parent +process into the child process and is mostly used to create new system context +for execve(2). vfork(2) calls are not counted towards kernel.all.sysfork. +@kernel.all.sysread Total number of system calls from read(2) family across all CPUs +@kernel.all.syswrite Total number of system calls from write (2) family across all CPUs + +@kernel.percpu.trap Traps on each CPUs +@kernel.percpu.pswitch Context switches on each CPUs +@kernel.percpu.syscall Total number of system calls on each CPU +@kernel.percpu.sysexec Total number of calls from exec(2) family on each CPU +@kernel.percpu.sysfork Total number of new processes created on each CPU +@kernel.percpu.sysvfork Total number of new processes created on each CPU +Unlike fork vfork does not copy all the virtual memory of the parent +process into the child process and is mostly used to create new system context +for execve(2). vfork(2) calls are not counted towards kernel.percpu.sysfork. +@kernel.percpu.sysread Total number of system calls from read(2) family on each CPU +@kernel.percpu.syswrite Total number of system calls from write (2) family on each CPU + +@kernel.all.load Classic load average for 1, 5 and 15 minute intervals + +@kernel.all.cpu.wait.io Time spent waiting for I/O across all CPUs +This metric is not updated by OpenSolaris kernel. +@kernel.all.cpu.wait.pio Time spent wait for polled I/O across all CPUs +This metric is not updated by OpenSolaris kernel. +@kernel.all.cpu.wait.swap Time spent wait for swap across all CPUs +This metric is not updated by OpenSolaris kernel. +@kernel.percpu.cpu.wait.io Time spent waiting for I/O on per-CPU basis +This metric is not updated by OpenSolaris kernel. +@kernel.percpu.cpu.wait.pio Time spent waiting for polled I/O on per-CPU basis +This metric is not updated by OpenSolaris kernel. +@kernel.percpu.cpu.wait.swap Time spent waiting swap on per-CPU basis +This metric is not updated by OpenSolaris kernel. + +@zfs.arc.size Total amount of memory used by ZFS ARC +@zfs.arc.min_size Lower limit of them amount of memory for ZFS ARC +@zfs.arc.max_size Upper limit of the amount of memory for ZFS ARC +The default is to use 7/8 of total physical memory. +@zfs.arc.mru_size Amount of memory used by the most recently used pages +@zfs.arc.target_size "Ideal" size of the cached based on aging +@zfs.arc.hits.total Number of times data is found in the cache +@zfs.arc.hits.mfu Number of times data is found in the most frequently used buffers +@zfs.arc.hits.mru Number of times data is found in the most recently used buffers +@zfs.arc.hits.mfu_ghost Number of times MFU ghost buffer is accessed +A ghost buffer is a buffer which is no longer cached but is still +linked into the hash. +@zfs.arc.hits.mru_ghost Number of times MRU ghost buffer is accessed +A ghost buffer is a buffer which is no longer cached but is still +linked into the hash. +@zfs.arc.hits.demand_data Number of times file data is found in the cache +ARC statistics provide separate counters for demand vs prefetch +and data vs metadata accesses: demand is a result of the direct +request for a particular data, prefetch is a result of speculative +request for a particular data. +@zfs.arc.hits.demand_metadata Number of times filesystem metadata is found in the cache +ARC statistics provide separate counters for demand vs prefetch +and data vs metadata accesses: demand is a result of the direct +request for a particular data, prefetch is a result of speculative +request for a particular data. +@zfs.arc.hits.prefetch_data Number of times speculative request for data is satisfied from the cache +ARC statistics provide separate counters for demand vs prefetch +and data vs metadata accesses: demand is a result of the direct +request for a particular data, prefetch is a result of speculative +request for a particular data. +@zfs.arc.hits.prefetch_metadata Number of times speculative request for metadata is satisfied from the cache +ARC statistics provide separate counters for demand vs prefetch +and data vs metadata accesses: demand is a result of the direct +request for a particular data, prefetch is a result of speculative +request for a particular data. +@zfs.arc.misses.total Number of times the data is not found in the cache +@zfs.arc.misses.demand_data Number of times file data is not found in the cache +ARC statistics provide separate counters for demand vs prefetch +and data vs metadata accesses: demand is a result of the direct +request for a particular data, prefetch is a result of speculative +request for a particular data. +@zfs.arc.misses.demand_metadata Number of times filesystem metadata is not found in the cache +ARC statistics provide separate counters for demand vs prefetch +and data vs metadata accesses: demand is a result of the direct +request for a particular data, prefetch is a result of speculative +request for a particular data. +@zfs.arc.misses.prefetch_data Number of times speculatively accessed file data is not found in the cache +ARC statistics provide separate counters for demand vs prefetch +and data vs metadata accesses: demand is a result of the direct +request for a particular data, prefetch is a result of speculative +request for a particular data. +@zfs.arc.misses.prefetch_metadata Number of times speculatively accessed filesystem metadata is not found in the cache +ARC statistics provide separate counters for demand vs prefetch +and data vs metadata accesses: demand is a result of the direct +request for a particular data, prefetch is a result of speculative +request for a particular data. +@pmda.prefetch.time Amount of time spent extracting information about a group of metrics +Each metric belongs to a prefetch group. When a client asks for the metric +to be fetched the information for the group must be extracted from the kernel. +@pmda.prefetch.count Number of times each group of metrics was updated + +@pmda.metric.time Amount of time spent extracting information about individual metrics +Requesting multiple instances of the same metrics counts against the metric +itself and not against the individual instances +@pmda.metric.count Number of times individual metrics have been fetched +Requesting multiple instances of the same metrics counts as multiple hits +against the metric itself + +@disk.all.wait.time Amount of time IO requests spent waiting for service +Amount of time IO transactions spent waiting to be serviced, i.e. the +transaction has been accepted for processing but for which the processing +has not yet begun. Each transaction waiting for processing adds to +to the total time which means that if multiple transactions are waiting then +total time for the sampling interval may be larger then the interval. + +@disk.dev.wait.time Amount of time IO requests spent waiting for service +Amount of time IO transactions spent waiting to be serviced, i.e. the +transaction has been accepted for processing but for which the processing +has not yet begun. Each transaction waiting for processing adds to +to the total time which means that if multiple transactions are waiting then +total time for the sampling interval may be larger then the interval. + +@disk.all.wait.count Number of transactions waiting to be serviced +Number of transactions accepted for processing but for which the processing +has not yet begun. +@disk.dev.wait.count Number of transactions waiting to be serviced +Number of transactions accepted for processing but for which the processing +has not yet begun. + +@disk.all.run.time Amount of time spent processing IO requests +@disk.dev.run.time Amount of time spent processing IO requests +@disk.all.run.count Number of transactions being processed +@disk.dev.run.count Number of transactions being processed + + +# from i86pc/os/cpuid.c +# /* +# * 8bit APIC IDs on dual core Pentiums +# * look like this: +# * +# * +-----------------------+------+------+ +# * | Physical Package ID | MC | HT | +# * +-----------------------+------+------+ +# * <------- chipid --------> +# * <------- coreid ---------------> +# * <--- clogid --> +# * <------> +# * pkgcoreid +# * +# * Where the number of bits necessary to +# * represent MC and HT fields together equals +# * to the minimum number of bits necessary to +# * store the value of cpi->cpi_ncpu_per_chip. +# * Of those bits, the MC part uses the number +# * of bits necessary to store the value of +# * cpi->cpi_ncore_per_chip. +# */ +# +@hinv.cpu.brand Marketing name of CPU +@hinv.cpu.clock Current CPU clock frequency +On CPUs which support dynamic clock frequency changes current clock frequency +could differ from the nominal ("maximum") clock frequency specified by +the manufacturer. +@hinv.cpu.maxclock Maximum clock frequency supported by CPU +Nominal CPU clock frequency as specified by the manufacturer. +@hinv.cpu.frequencies List of clock frequencies supported by CPU +@hinv.cpu.implementation Details of CPU implementation +@hinv.cpu.chip_id Chip or Socket identifier of the CPU +Logical CPUs can share single chip identifier +@hinv.cpu.clog_id Logical core identifier +Logical cores identifier combines identifiers of the CPU core and +virtual CPU identifier (aka hyperthread identifier). +@hinv.cpu.core_id CPU core identifier +CPU core identifire combines chip identifier and per-chip core +identifier. If cores support more the one virtual CPU per core +then same core identifier is shared across several virtual +CPUs. +@hinv.cpu.pkg_core_id Per-chip core identifier +This identifier is used to identify individual cores within the +package. If a core support more the one virtual CPU per core +then same core identifier is shared across several virtual +CPUs. +@hinv.cpu.cstate Current CPU idle state +@hinv.cpu.maxcstates Maximum number of idle state supported by the CPU +Information about cstate is available in kstat(1m). +@hinv.cpu.ncores Number of CPU cores per physical chip +@hinv.cpu.ncpus Number of virtual CPUs per physical chip + +@disk.dev.errors.soft Number of soft errors per device +@disk.dev.errors.hard Number of hard errors per device +@disk.dev.errors.transport Number of transport errors per device +@disk.dev.errors.media Number of media errors per device +@disk.dev.errors.recoverable Number of recoverable errors per device +@disk.dev.errors.notready Number of times device reported as not ready +@disk.dev.errors.nodevice Number of times device was found missing +@disk.dev.errors.badrequest Number of illegal requests per device +@disk.dev.errors.pfa Number of times failure prediction threshold has been exceeded +@hinv.disk.vendor Device Vendor +Can be reported as ATA if SATA device is behind SAS expander +@hinv.disk.product Device name +Vendor's device name (up-to 16 characters long) +@hinv.disk.revision Device Revision +@hinv.disk.serial Device Serial Number +@hinv.disk.capacity Device Capacity +For removable devices capacity of the media is reported. + +@kernel.fs.vnops.access Number of VOP_ACCESS calls per filesystem +VOP_ACCESS is used by access(2) system call. +@kernel.fs.vnops.addmap Number of VOP_ADDMAP calls per filesystem +VOP_ADDMAP is used to manage reference counting of the vnode used by +mmap(2) operations. +@kernel.fs.vnops.close Number of VOP_CLOSE calls per filesystem +VOP_CLOSE is called every time a close(2) system call is called +@kernel.fs.vnops.cmp Number of VOP_CMP calls per filesystem +VOP_CMP is used to check if two vnodes are "equal" to each other, i.e. +both refer to the same filesystem object. +@kernel.fs.vnops.create Number of VOP_CREATE calls per filesystem +VOP_CREATE is used to create regular files and device or FIFO nodes. +@kernel.fs.vnops.delmap Number of VOP_DELMAP calls +VOP_DELMAP is used to destroy a previously created memory-mapped region +of a file. +@kernel.fs.vnops.dispose Number ot VOP_DISPOSE calls per filesystem +VOP_DISPOSE is used to dispose(free or invalidate) of a page associated +with a file. +@kernel.fs.vnops.dump Number of VOP_DUMP calls per filesystem +VOP_DUMP is used to transfer data from the frozen kernel directly +to the dump device +@kernel.fs.vnops.dumpctl Number of VOP_DUMPCTL calls per filesystem +VOP_DUMPCTL sets up context used by VOP_DUMP call. It is used to +allocate, free or search for data blocks on the dump device. +@kernel.fs.vnops.fid Number of VOP_FID calls per filesystem +VOP_FID is used to get file identifier which can be used instead of the +file name in some operations. NFS server is one known user of this vnode +operation. +@kernel.fs.vnops.frlock Number of VOP_FRLOCK calls per filesystem +VOP_FRLOCK is used to implement file record locking used by flock(2) +@kernel.fs.vnops.fsync Number of VOP_FSYNC calls per filesystem +VOP_FSYNC is used to implement fsync(2) system call which flushes +data for a specific file to disk. +@kernel.fs.vnops.getattr Number of VOP_GETATTR calls per filesystem +VOP_GETATTR is used to extract vnode attributes. It use used as part of +many system calls which manipulate file attributes, e.g. chmod(2), stat(2), +utimes(2) etc. +@kernel.fs.vnops.getpage Number of VOP_GETPAGE calls per filesystem +VOP_GETPAGE is used to allocate pages (could be several at a time) to cover +a region in a file. +@kernel.fs.vnops.getsecattr Number of VOP_GETSECATTR calls per filesystem +VOP_GETSECATTR used to extract ACL entires associated with a file. +@kernel.fs.vnops.inactive Number of VOP_INACTIVE calls per filesystem +VOP_INACTIVE is used to destroy vnode before it is removed from the +cache or reused. +@kernel.fs.vnops.ioctl Number of VOP_IOCTL calls per filesystem +VOP_IOCTL is used to implement ioctl(2) system call. +@kernel.fs.vnops.link Number of VOP_LINK calls per filesystem +VOP_LINK is used to implement support for hard links +@kernel.fs.vnops.lookup Number of VOP_LOOKUP calls per filesystem +VOP_LOOKUP is used to translate filename to vnode. +@kernel.fs.vnops.map Number of VOP_MAP calls per filesystem +VOP_MAP is used to create a new memory-mapped region of a file +@kernel.fs.vnops.mkdir Number of VOP_MKDIR calls per filesystem +VOP_MKDIR is used to create directories +@kernel.fs.vnops.open Number of VOP_OPEN calls per filesystem +VOP_OPEN is called every time open(2) system call is called. +@kernel.fs.vnops.pageio Number of VOP_PAGEIO calls per filesystem +VOP_PAGEIO is similar to VOP_GETPAGE and VOP_PUTPAGE and can be used when +either of the other two are less efficient, e.g. in the case when pages +will be reused after the IO is done. +@kernel.fs.vnops.pathconf Number of VOP_PATHCONF calls per filesystem +VOP_PATHCONF is used to obtain information about filesystem's parameters +reported by pathconf(2) system call +@kernel.fs.vnops.poll Number of VOP_POLL calls per filesystem +VOP_POLL is used to implement pool(2) system call +@kernel.fs.vnops.putpage Number of VOP_PUTPAGE calls per filesystem +VOP_PUTPAGE is used to release pages which have been used to hold +data from a file +@kernel.fs.vnops.read Number of VOP_READ calls per filesystem +VOP_READ is used to implement read(2) system call +@kernel.fs.vnops.readdir Number of VOP_READDIR calls per filesystem +VOP_READDIR is used to read directory entries +@kernel.fs.vnops.readlink Number of VOP_READLINK calls per filesystem +VOP_READLINK is used to read the information about the target of the symbolic +link +@kernel.fs.vnops.realvp Number of VOP_REALVP calls per filesystem +VOP_REALVP is used to traverse stacking filesystems and extract information +about the vnode which refers to the "real" filesystem object. +@kernel.fs.vnops.remove Number of VOP_REMOVE calls per filesystem +VOP_REMOVE is used to remove entires from a directory. +@kernel.fs.vnops.rename Number of VOP_RENAME calls per filesystem +VOP_RENAME is used to implement rename(2) system call +@kernel.fs.vnops.rmdir Number of VOP_RMDIR calls per filesystem +VOP_RMDIR is used to implement rmdir(2) system call +@kernel.fs.vnops.rwlock Number of VOP_RWLOCK calls per filesystem +VOP_RWLOCK and VOP_RWUNLOCK are used to protect access to vnode data. +@kernel.fs.vnops.rwunlock Number of VOP_RWUNLOCK calls per filesystem +VOP_RWLOCK and VOP_RWUNLOCK are used to protect access to vnode data. +@kernel.fs.vnops.seek Number of VOP_SEEK calls per filesystem +VOP_SEEK is used by lseek(2). Because vnodes can be shared across multiple +instances of vfile VOP_SEEK does not usually change the position of the +file pointer, it instead used to verify the offset before it is changed. +@kernel.fs.vnops.setattr Number of VOP_SETATTR calls per filesystem +VOP_SETATTR is used to change vnode attributes which are modified by system +calls like chmod(2), chown(2), utimes(2) etc. +@kernel.fs.vnops.setfl Number of VOP_SETFL calls per filesystem +VOP_SETFL is used to implement fcntl(2) F_SETFL option. +Currently only sockfs pseudo filesystem is implementing this vnode operation. +@kernel.fs.vnops.setsecattr Number of VOP_SETSECATTR calls per filesystem +VOP_SETSECATTR is used to change ACL entries +@kernel.fs.vnops.shrlock Number of VOP_SHRLOCK calls per filesystem +VOP_SHRLOCK is usually used to implement CIFS and NLMv3 shared reservations. +@kernel.fs.vnops.space Number of VOP_SPACE calls per filesystem +VOP_SPACE is used to provide optimized support for growing and shrinking the files. +F_FREESP option of fcntl(2) is using this vnode operation to implment ftruncate(3c) +function. +@kernel.fs.vnops.symlink Number of VOP_SYMLINK calls per filesystem +VOP_SYMLINK is used to create symbolic links. +@kernel.fs.vnops.vnevent Number of VOP_VNEVENT calls per filesystem +VIP_VNEVENT is used to check if a filesystem support vnode event +notifications for operations which change the names of the files. +@kernel.fs.vnops.write Number of VOP_WRITE calls per filesystem +VOP_WRITE is used to implement write(2) system call +@kernel.fs.read_bytes Number of bytes read from a specific filesystem +@kernel.fs.readdir_bytes Number of bytes containting directory entires read from a specific filesystem +@kernel.fs.write_bytes Number of bytes written to a specific filesystem + +@kernel.fstype.vnops.access Number of VOP_ACCESS calls for all filesystems of a given type +VOP_ACCESS is used by access(2) system call. +@kernel.fstype.vnops.addmap Number of VOP_ADDMAP calls for all filesystems of a given type +VOP_ADDMAP is used to manage reference counting of the vnode used by +mmap(2) operations. +@kernel.fstype.vnops.close Number of VOP_CLOSE calls for the specific filesystem +VOP_CLOSE is called every time a close(2) system call is called +@kernel.fstype.vnops.cmp Number of VOP_CMP calls for the specific filesystem +VOP_CMP is used to check if two vnodes are "equal" to each other, i.e. +both refer to the same filesystem object. +@kernel.fstype.vnops.create Number of VOP_CREATE calls for all filesystems of a given type +VOP_CREATE is used to create regular files and device or FIFO nodes. +@kernel.fstype.vnops.delmap Number of VOP_DELMAP was called +VOP_DELMAP is used to destroy a previously created memory-mapped region +of a file. +@kernel.fstype.vnops.dispose Number ot VOP_DISPOSE calls for all filesystems of a given type +VOP_DISPOSE is used to dispose(free or invalidate) of a page associated +with a file. +@kernel.fstype.vnops.dump Number of VOP_DUMP calls for all filesystems of a given type +VOP_DUMP is used to transfer data from the frozen kernel directly +to the dump device +@kernel.fstype.vnops.dumpctl Number of VOP_DUMPCTL calls for all filesystems of a given type +VOP_DUMPCTL sets up context used by VOP_DUMP call. It is used to +allocate, free or search for data blocks on the dump device. +@kernel.fstype.vnops.fid Number of VOP_FID calls for all filesystems of a given type +VOP_FID is used to get file identifier which can be used instead of the +file name in some operations. NFS server is one known user of this vnode +operation. +@kernel.fstype.vnops.frlock Number of time VOP_FRLOCK calls for all filesystems of a given type +VOP_FRLOCK is used to implement file record locking used by flock(2) +@kernel.fstype.vnops.fsync Number of VOP_FSYNC calls for all filesystems of a given type +VOP_FSYNC is used to implement fsync(2) system call which flushes +data for a specific file to disk. +@kernel.fstype.vnops.getattr Number of VOP_GETATTR calls for all filesystems of a given type +VOP_GETATTR is used to extract vnode attributes. It use used as part of +many system calls which manipulate file attributes, e.g. chmod(2), stat(2), +utimes(2) etc. +@kernel.fstype.vnops.getpage Number of VOP_GETPAGE calls for all filesystems of a given type +VOP_GETPAGE is used to allocate pages (could be several at a time) to cover +a region in a file. +@kernel.fstype.vnops.getsecattr Number of VOP_GETSECATTR calls for all filesystems of a given type +VOP_GETSECATTR used to extract ACL entires associated with a file. +@kernel.fstype.vnops.inactive Number of VOP_INACTIVE calls for all filesystems of a given type +VOP_INACTIVE is used to destroy vnode before it is removed from the +cache or reused. +@kernel.fstype.vnops.ioctl Number of VOP_IOCTL calls for all filesystems of a given type +VOP_IOCTL is used to implement ioctl(2) system call. +@kernel.fstype.vnops.link Number of VOP_LINK calls for all filesystems of a given type +VOP_LINK is used to implement support for hard links +@kernel.fstype.vnops.lookup Number of VOP_LOOKUP calls for all filesystems of a given type +VOP_LOOKUP is used to translate filename to vnode. +@kernel.fstype.vnops.map Number of VOP_MAP calls for all filesystems of a given type +VOP_MAP is used to create a new memory-mapped region of a file +@kernel.fstype.vnops.mkdir Number of VOP_MKDIR calls for all filesystems of a given type +VOP_MKDIR is used to create directories +@kernel.fstype.vnops.open Number of VOP_OPEN calls for all filesystems of a given type +VOP_OPEN is called every time open(2) system call is called. +@kernel.fstype.vnops.pageio Number of VOP_PAGEIO calls for all filesystems of a given type +VOP_PAGEIO is similar to VOP_GETPAGE and VOP_PUTPAGE and can be used when +either of the other two are less efficient, e.g. in the case when pages +will be reused after the IO is done. +@kernel.fstype.vnops.pathconf Number of VOP_PATHCONF calls for all filesystems of a given type +VOP_PATHCONF is used to obtain information about filesystem's parameters +reported by pathconf(2) system call +@kernel.fstype.vnops.poll Number of VOP_POLL calls for all filesystems of a given type +VOP_POLL is used to implement pool(2) system call +@kernel.fstype.vnops.putpage Number of VOP_PUTPAGE calls for all filesystems of a given type +VOP_PUTPAGE is used to release pages which have been used to hold +data from a file +@kernel.fstype.vnops.read Number of VOP_READ calls for all filesystems of a given type +VOP_READ is used to implement read(2) system call +@kernel.fstype.vnops.readdir Number of VOP_READDIR calls for all filesystems of a given type +VOP_READDIR is used to read directory entries +@kernel.fstype.vnops.readlink Number of VOP_READLINK calls for all filesystems of a given type +VOP_READLINK is used to read the information about the target of the symbolic +link +@kernel.fstype.vnops.realvp Number of VOP_REALVP calls for all filesystems of a given type +VOP_REALVP is used to traverse stacking filesystems and extract information +about the vnode which refers to the "real" filesystem object. +@kernel.fstype.vnops.remove Number of VOP_REMOVE calls for all filesystems of a given type +VOP_REMOVE is used to remove entires from a directory. +@kernel.fstype.vnops.rename Number of VOP_RENAME calls for all filesystems of a given type +VOP_RENAME is used to implement rename(2) system call +@kernel.fstype.vnops.rmdir Number of VOP_RMDIR calls for all filesystems of a given type +VOP_RMDIR is used to implement rmdir(2) system call +@kernel.fstype.vnops.rwlock Number of VOP_RWLOCK calls for all filesystems of a given type +VOP_RWLOCK and VOP_RWUNLOCK are used to protect access to vnode data. +@kernel.fstype.vnops.rwunlock Number of VOP_RWUNLOCK calls for all filesystems of a given type +VOP_RWLOCK and VOP_RWUNLOCK are used to protect access to vnode data. +@kernel.fstype.vnops.seek Number of VOP_SEEK calls for all filesystems of a given type +VOP_SEEK is used by lseek(2). Because vnodes can be shared across multiple +instances of vfile VOP_SEEK does not usually change the position of the +file pointer, it instead used to verify the offset before it is changed. +@kernel.fstype.vnops.setattr Number of VOP_SETATTR calls for all filesystems of a given type +VOP_SETATTR is used to change vnode attributes which are modified by system +calls like chmod(2), chown(2), utimes(2) etc. +@kernel.fstype.vnops.setfl Number of VOP_SETFL calls for all filesystems of a given type +VOP_SETFL is used to implement fcntl(2) F_SETFL option. +Currently only sockfs pseudo filesystem is implementing this vnode operation. +@kernel.fstype.vnops.setsecattr Number of VOP_SETSECATTR calls for all filesystems of a given type +VOP_SETSECATTR is used to change ACL entries +@kernel.fstype.vnops.shrlock Number of VOP_SHRLOCK calls for all filesystems of a given type +VOP_SHRLOCK is usually used to implement CIFS and NLMv3 shared reservations. +@kernel.fstype.vnops.space Number of VOP_SPACE calls for all filesystems of a given type +VOP_SPACE is used to provide optimized support for growing and shrinking the files. +F_FREESP option of fcntl(2) is using this vnode operation to implment ftruncate(3c) +function. +@kernel.fstype.vnops.symlink Number of VOP_SYMLINK calls for all filesystems of a given type +VOP_SYMLINK is used to create symbolic links. +@kernel.fstype.vnops.vnevent Number of VOP_VNEVENT calls for all filesystems of a given type +VIP_VNEVENT is used to check if a filesystem support vnode event +notifications for operations which change the names of the files. +@kernel.fstype.vnops.write Number of VOP_WRITE calls for all filesystems of a given type +VOP_WRITE is used to implement write(2) system call +@kernel.fstype.read_bytes Bytes read from all filesystems of a given type +@kernel.fstype.readdir_bytes Bytes read for directory entries from all filesystems of a given type +@kernel.fstype.write_bytes Bytes written to all filesystems of a given type + +@hinv.disk.devlink Disk name in the descriptive format +Solaris uses symbolic links under /dev to provide access to device nodes via +"descriptive" names like /dev/dsk/cXtYdZsN. This metrics provides a +translation from a "descriptive" name to instances in the disk instance +domain. + +The name is always the name of the first minor device for a particular disk +and includes the slice information. + +NOTE! Fetching this metric is expensive - several system calls are made + to fetch each instance. |