{ "content": [ { "type": "text", "text": "# MULTIPATH.CONF (man)\n\n## NAME\n\nmultipath.conf - multipath daemon configuration file.\n\n## DESCRIPTION\n\n/etc/multipath.conf is the configuration file for the multipath daemon. It is used to over‐\nwrite the built-in configuration table of multipathd. Any line whose first non-white-space\ncharacter is a '#' is considered a comment line. Empty lines are ignored.\n\n## Sections\n\n- **NAME**\n- **DESCRIPTION**\n- **SYNTAX** (6 subsections)\n- **FOREIGN MULTIPATH SUPPORT**\n- **KNOWN ISSUES**\n- **SEE ALSO**\n- **AUTHORS**\n\nUse structuredContent.sections for detailed options, examples, and full documentation.\n" } ], "structuredContent": { "command": "MULTIPATH.CONF", "section": "", "mode": "man", "summary": "multipath.conf - multipath daemon configuration file.", "synopsis": null, "tldr_summary": null, "tldr_examples": [], "tldr_source": null, "flags": [], "examples": [], "see_also": [ { "name": "udev", "section": "8", "url": "https://www.chedong.com/phpMan.php/man/udev/8/json" }, { "name": "dmsetup", "section": "8", "url": "https://www.chedong.com/phpMan.php/man/dmsetup/8/json" }, { "name": "multipath", "section": "8", "url": "https://www.chedong.com/phpMan.php/man/multipath/8/json" }, { "name": "multipathd", "section": "8", "url": "https://www.chedong.com/phpMan.php/man/multipathd/8/json" } ], "section_outline": [ { "name": "NAME", "lines": 2, "subsections": [] }, { "name": "DESCRIPTION", "lines": 7, "subsections": [] }, { "name": "SYNTAX", "lines": 56, "subsections": [ { "name": "defaults section", "lines": 850 }, { "name": "multipaths section", "lines": 48 }, { "name": "devices section", "lines": 107 }, { "name": "overrides section", "lines": 39 }, { "name": "WWID generation", "lines": 26 }, { "name": "Shaky paths detection", "lines": 47 } ] }, { "name": "FOREIGN MULTIPATH SUPPORT", "lines": 16, "subsections": [] }, { "name": "KNOWN ISSUES", "lines": 9, "subsections": [] }, { "name": "SEE ALSO", "lines": 2, "subsections": [] }, { "name": "AUTHORS", "lines": 6, "subsections": [] } ], "sections": { "NAME": { "content": "multipath.conf - multipath daemon configuration file.\n", "subsections": [] }, "DESCRIPTION": { "content": "/etc/multipath.conf is the configuration file for the multipath daemon. It is used to over‐\nwrite the built-in configuration table of multipathd. Any line whose first non-white-space\ncharacter is a '#' is considered a comment line. Empty lines are ignored.\n\nCurrently used multipathd configuration can be displayed with the multipath -t or multipathd\nshow config command.\n", "subsections": [] }, "SYNTAX": { "content": "The configuration file contains entries of the form:\n\n
{\n \n...\n {\n \n...\n}\n}\n\nEach section contains one or more attributes or subsections. The recognized keywords for at‐\ntributes or subsections depend on the section in which they occur.\n\n and must be on a single line. is one of the keywords listed\nin this man page. is either a simple word (containing no whitespace and none of the\ncharacters '\"', '#', and '!') or one string enclosed in double quotes (\"...\"). Outside a\nquoted string, text starting with '#', and '!' is regarded as a comment and ignored until the\nend of the line. Inside a quoted string, '#' and '!' are normal characters, and whitespace is\npreserved. To represent a double quote character inside a double quoted string, use two con‐\nsecutive double quotes ('\"\"'). Thus '2.5\" SSD' can be written as \"2.5\"\" SSD\".\n\nOpening braces ('{') must follow the (sub)section name on the same line. Closing braces ('}')\nthat mark the end of a (sub)section must be the only non-whitespace character on the line.\nWhitespace is ignored except inside double quotes, thus the indentation shown in the above\nexample is helpful for human readers but not mandatory.\n\nNote on regular expressions: The multipath.conf syntax allows many attribute values to be\nspecified as POSIX Extended Regular Expressions (see regex(7)). These regular expressions are\ncase sensitive and not anchored, thus the expression \"bar\" matches \"barbie\", \"rhabarber\", and\n\"wunderbar\", but not \"Barbie\". To avoid unwanted substring matches, standard regular expres‐\nsion syntax using the special characters \"^\" and \"$\" can be used.\n\nThe following section keywords are recognized:\n\ndefaults This section defines default values for attributes which are used whenever\nno values are given in the appropriate device or multipath sections.\n\nblacklist This section defines which devices should be excluded from the multipath\ntopology discovery.\n\nblacklistexceptions\nThis section defines which devices should be included in the multipath\ntopology discovery, despite being listed in the blacklist section.\n\nmultipaths This section defines the multipath topologies. They are indexed by a World\nWide Identifier(WWID). For details on the WWID generation see section WWID\ngeneration below. Attributes set in this section take precedence over all\nothers.\n\ndevices This section defines the device-specific settings. Devices are identified by\nvendor, product, and revision.\n\noverrides This section defines values for attributes that should override the device-\nspecific settings for all devices.\n", "subsections": [ { "name": "defaults section", "content": "The defaults section recognizes the following keywords:\n\nverbosity Default verbosity. Higher values increase the verbosity level. Valid levels\nare between 0 and 6.\n\nThe default is: 2\n\npollinginterval Interval between two path checks in seconds. For properly functioning paths,\nthe interval between checks will gradually increase to maxpollinginterval.\nThis value will be overridden by the WatchdogSec setting in the multi‐\npathd.service definition if systemd is used.\n\nThe default is: 5\n\nmaxpollinginterval\nMaximal interval between two path checks in seconds.\n\nThe default is: 4 * pollinginterval\n\nreassignmaps Enable reassigning of device-mapper maps. With this option multipathd will\nremap existing device-mapper maps to always point to multipath device, not\nthe underlying block devices. Possible values are yes and no.\n\nThe default is: no\n\nmultipathdir This option is deprecated, and will be removed in a future release. Direc‐\ntory where the dynamic shared objects are stored. Defined at compile time,\ncommonly /lib64/multipath/ or /lib/multipath/.\n\nThe default is: \n\npathselector The default path selector algorithm to use; they are offered by the kernel\nmultipath target:\n\nround-robin 0\nLoop through every path in the path group, sending the same\namount of I/O to each. Some aspects of behavior can be con‐\ntrolled with the attributes: rrminio, rrminiorq and\nrrweight.\n\nqueue-length 0\n(Since 2.6.31 kernel) Choose the path for the next bunch of I/O\nbased on the amount of outstanding I/O to the path.\n\nservice-time 0\n(Since 2.6.31 kernel) Choose the path for the next bunch of I/O\nbased on the amount of outstanding I/O to the path and its rela‐\ntive throughput.\n\nhistorical-service-time 0\n(Since 5.8 kernel) Choose the path for the next bunch of I/O\nbased on the estimation of future service time based on the his‐\ntory of previous I/O submitted to each path.\n\nio-affinity 0\n(Since 5.11 kernel) Choose the path for the next bunch of I/O\nbased on a CPU to path mapping the user passes in and what CPU\nwe are executing on.\n\nThe default is: service-time 0\n\npathgroupingpolicy\nThe default path grouping policy to apply to unspecified multipaths. Possi‐\nble values are:\n\nfailover One path per priority group.\n\nmultibus All paths in one priority group.\n\ngroupbyserial\nOne priority group per serial number.\n\ngroupbyprio\nOne priority group per priority value. Priorities are determined\nby callout programs specified as a global, per-controller or\nper-multipath option in the configuration file.\n\ngroupbynodename\nOne priority group per target node name. Target node names are\nfetched in /sys/class/fctransport/target*/nodename.\n\nThe default is: failover\n\nuidattrs Setting this option activates merging uevents by WWID, which may improve\nuevent processing effiency. Moreover, it's an alternative method to config‐\nure the udev properties to use for determining unique path identifiers\n(WWIDs).\n\nThe value of this option is a space separated list of records like\n\"type:ATTR\", where type is matched against the beginning of the device node\nname (e.g. sd:ATTR matches sda), and ATTR is the name of the udev property\nto use for matching devices.\n\nIf this option is configured and matches the device node name of a device,\nit overrides any other configured methods for determining the WWID for this\ndevice.\n\nThe default is: . To enable uevent merging, set it e.g. to \"sd:IDSE‐\nRIAL dasd:IDUID nvme:IDWWN\".\n\nuidattribute The udev attribute providing a unique path identifier (WWID). If uidattri‐\nbute is set to the empty string, WWID determination is done using the sysfs\nmethod rather then using udev (not recommended in production; see WWID gen‐‐\neration below).\n\nThe default is: IDSERIAL, for SCSI devices\n\nThe default is: IDUID, for DASD devices\n\nThe default is: IDWWN, for NVMe devices\n\ngetuidcallout (Superseded by uidattribute) The default program and args to callout to ob‐\ntain a unique path identifier. Should be specified with an absolute path.\n\nThe default is: \n\nprio The name of the path priority routine. The specified routine should return a\nnumeric value specifying the relative priority of this path. Higher number\nhave a higher priority. \"none\" is a valid value. Currently the following\npath priority routines are implemented:\n\nconst Return a constant priority of 1.\n\nsysfs Use the sysfs attributes accessstate and preferredpath to gen‐\nerate the path priority. This prioritizer accepts the optional\nprioarg exclusiveprefbit.\n\nemc (Hardware-dependent) Generate the path priority for DGC class\narrays as CLARiiON CX/AX and EMC VNX and Unity families.\n\nalua (Hardware-dependent) Generate the path priority based on the\nSCSI-3 ALUA settings. This prioritizer accepts the optional\nprioarg exclusiveprefbit.\n\nontap (Hardware-dependent) Generate the path priority for NetApp ONTAP\nclass and OEM arrays as IBM NSeries.\n\nrdac (Hardware-dependent) Generate the path priority for LSI/Enge‐\nnio/NetApp RDAC class as NetApp SANtricity E/EF Series, and OEM\narrays from IBM DELL SGI STK and SUN.\n\nhpsw (Hardware-dependent) Generate the path priority for HP/COM‐\nPAQ/DEC HSG80 and MSA/HSV arrays with Active/Standby mode exclu‐\nsively.\n\nhds (Hardware-dependent) Generate the path priority for Hitachi AMS\nfamilies of arrays other than AMS 2000.\n\nrandom Generate a random priority between 1 and 10.\n\nweightedpath\nGenerate the path priority based on the regular expression and\nthe priority provided as argument. Requires prioargs keyword.\n\npathlatency\nGenerate the path priority based on a latency algorithm. Re‐\nquires prioargs keyword.\n\nana (Hardware-dependent) Generate the path priority based on the\nNVMe ANA settings.\n\ndatacore (Hardware-dependent) Generate the path priority for some Data‐\nCore storage arrays. Requires prioargs keyword.\n\niet (iSCSI only) Generate path priority for iSCSI targets based on\nIP address. Requires prioargs keyword.\n\nThe default depends on the detectprio setting: If detectprio is yes (de‐\nfault), the default priority algorithm is sysfs (except for NetAPP E-Series,\nwhere it is alua). If detectprio is no, the default priority algorithm is\nconst.\n\nprioargs Arguments to pass to to the prio function. This only applies to certain pri‐\noritizers:\n\nweighted Needs a value of the form \" \n ...\"\n\nhbtl Regex can be of SCSI H:B:T:L format. For example:\n1:0:.:. , *:0:0:.\n\ndevname Regex can be of device name format. For example: sda ,\nsd.e\n\nserial Regex can be of serial number format. For example:\n.*J1FR.*324 . The serial can be looked up through sysfs\nor by running multipathd show paths format \"%z\". For ex‐\nample: 0395J1FR904324\n\nwwn Regex can be of the form \"hostwwnn:hostwwpn:tar‐\ngetwwnn:targetwwpn\" these values can be looked up\nthrough sysfs or by running multipathd show paths format\n\"%N:%R:%n:%r\". For example:\n0x200100e08ba0aea0:0x210100e08ba0aea0:.*:.* ,\n.*:.*:iqn.2009-10.com.redhat.msp.lab.ask-06:.*\n\npathlatency\nNeeds a value of the form \"ionum=<20> basenum=<10>\"\n\nionum The number of read IOs sent to the current path continu‐\nously, used to calculate the average path latency.\nValid Values: Integer, [2, 200].\n\nbasenum\nThe base number value of logarithmic scale, used to par‐\ntition different priority ranks. Valid Values: Integer,\n[2, 10]. And Max average latency value is 100s, min av‐\nerage latency value is 1us. For example: If\nbasenum=10, the paths will be grouped in priority\ngroups with path latency <=1us, (1us, 10us], (10us,\n100us], (100us, 1ms], (1ms, 10ms], (10ms, 100ms],\n(100ms, 1s], (1s, 10s], (10s, 100s], >100s.\n\nalua If exclusiveprefbit is set, paths with the preferred path bit\nset will always be in their own path group.\n\nsysfs If exclusiveprefbit is set, paths with the preferred path bit\nset will always be in their own path group.\n\ndatacore\n\npreferredsds\n(Mandatory) The preferred \"SDS name\".\n\ntimeout (Optional) The timeout for the INQUIRY, in ms.\n\niet\n\npreferredip=...\n(Mandatory) Th preferred IP address, in dotted decimal\nnotation, for iSCSI targets.\n\nThe default is: \n\nfeatures Specify any device-mapper features to be used. Syntax is num list where num\nis the number, between 0 and 8, of features in list. Possible values for\nthe feature list are:\n\nqueueifnopath\n(Deprecated, superseded by nopathretry) Queue I/O if no path\nis active. Identical to the nopathretry with queue value. If\nboth this feature and nopathretry are set, the latter value\ntakes precedence. See KNOWN ISSUES.\n\npginitretries \n(Since kernel 2.6.24) Number of times to retry pginit, it must\nbe between 1 and 50.\n\npginitdelaymsecs \n(Since kernel 2.6.38) Number of msecs before pginit retry, it\nmust be between 0 and 60000.\n\nqueuemode \n(Since kernel 4.8) Select the the queueing mode per multipath\ndevice. can be bio, rq or mq, which corresponds to bio-\nbased, request-based, and block-multiqueue (blk-mq) request-\nbased, respectively. The default depends on the kernel parame‐\nter dmmod.useblkmq. It is mq if the latter is set, and rq\notherwise.\n\nThe default is: \n\npathchecker The default method used to determine the path's state. The synchronous\ncheckers (all except tur and directio) will cause multipathd to pause most\nactivity, waiting up to checkertimeout seconds for the path to respond. The\nasynchronous checkers (tur and directio) will not pause multipathd. Instead,\nmultipathd will check for a response once per second, until checkertimeout\nseconds have elapsed. Possible values are:\n\nreadsector0 (Deprecated) Read the first sector of the device. This checker\nis being deprecated, please use tur instead.\n\ntur Issue a TEST UNIT READY command to the device.\n\nemcclariion\n(Hardware-dependent) Query the DGC/EMC specific EVPD page 0xC0\nto determine the path state for CLARiiON CX/AX and EMC VNX and\nUnity arrays families.\n\nhpsw (Hardware-dependent) Check the path state for HP/COMPAQ/DEC\nHSG80 and MSA/HSV arrays with Active/Standby mode exclusively.\n\nrdac (Hardware-dependent) Check the path state for LSI/Engenio/NetApp\nRDAC class as NetApp SANtricity E/EF Series, and OEM arrays from\nIBM DELL SGI STK and SUN.\n\ndirectio Read the first sector with direct I/O. This checker could cause\nspurious path failures under high load. Increasing checkertime‐\nout can help with this.\n\nccisstur (Hardware-dependent) Check the path state for HP/COMPAQ Smart\nArray(CCISS) controllers.\n\nnone Do not check the device, fallback to use the values retrieved\nfrom sysfs\n\nThe default is: tur\n\naliasprefix The userfriendlynames prefix.\n\nThe default is: mpath\n\nfailback Tell multipathd how to manage path group failback. To select immediate or a\nvalue, it's mandatory that the device has support for a working prioritizer.\n\nimmediate Immediately failback to the highest priority pathgroup that con‐\ntains active paths.\n\nmanual Do not perform automatic failback.\n\nfollowover Used to deal with multiple computers accessing the same Ac‐\ntive/Passive storage devices. Only perform automatic failback\nwhen the first path of a pathgroup becomes active. This keeps a\ncluster node from automatically failing back when another node\nrequested the failover.\n\nvalues > 0 Deferred failback (time to defer in seconds).\n\nThe default is: manual\n\nrrminio Number of I/O requests to route to a path before switching to the next in\nthe same path group. This is only for Block I/O(BIO) based multipath and\nonly apply to round-robin pathselector.\n\nThe default is: 1000\n\nrrminiorq Number of I/O requests to route to a path before switching to the next in\nthe same path group. This is only for Request based multipath and only apply\nto round-robin pathselector.\n\nThe default is: 1\n\nmaxfds Specify the maximum number of file descriptors that can be opened by multi‐\npath and multipathd. This is equivalent to ulimit -n. A value of max will\nset this to the system limit from /proc/sys/fs/nropen. If this is not set,\nthe maximum number of open fds is taken from the calling process. It is usu‐\nally 1024. To be safe, this should be set to the maximum number of paths\nplus 32, if that number is greated than 1024.\n\nThe default is: max\n\nrrweight If set to priorities the multipath configurator will assign path weights as\n\"path prio * rrminio\". Possible values are priorities or uniform . Only\napply to round-robin pathselector.\n\nThe default is: uniform\n\nnopathretry Specify what to do when all paths are down. Possible values are:\n\nvalue > 0 Number of retries until disable I/O queueing.\n\nfail For immediate failure (no I/O queueing).\n\nqueue For never stop I/O queueing, similar to queueifnopath. See\nKNOWN ISSUES.\n\nThe default is: fail\n\nqueuewithoutdaemon\nIf set to no , when multipathd stops, queueing will be turned off for all\ndevices. This is useful for devices that set nopathretry. If a machine is\nshut down while all paths to a device are down, it is possible to hang wait‐\ning for I/O to return from the device after multipathd has been stopped.\nWithout multipathd running, access to the paths cannot be restored, and the\nkernel cannot be told to stop queueing I/O. Setting queuewithoutdaemon to\nno , avoids this problem.\n\nThe default is: no\n\ncheckertimeout Specify the timeout to use for path checkers and prioritizers, in seconds.\nOnly prioritizers that issue scsi commands use checkertimeout. If a path\ndoes not respond to the checker command after checkertimeout seconds have\nelapsed, it is considered down.\n\nThe default is: in /sys/block//device/timeout\n\nallowusbdevices\nIf set to no , all USB devices will be skipped during path discovery. If you\nintend to use multipath on USB attached devices, set this to yes.\n\nThe default is: no\n\nflushonlastdel\nIf set to yes , multipathd will disable queueing when the last path to a de‐\nvice has been deleted.\n\nThe default is: no\n\nuserfriendlynames\nIf set to yes , using the bindings file /etc/multipath/bindings to assign a\npersistent and unique alias to the multipath, in the form of mpath. If\nset to no use the WWID as the alias. In either case this be will be overrid‐\nden by any specific aliases in the multipaths section.\n\nThe default is: no\n\nfastiofailtmo Specify the number of seconds the SCSI layer will wait after a problem has\nbeen detected on a FC remote port before failing I/O to devices on that re‐\nmote port. This should be smaller than devlosstmo. Setting this to off\nwill disable the timeout.\n\nThe default is: 5\n\ndevlosstmo Specify the number of seconds the SCSI layer will wait after a problem has\nbeen detected on a FC remote port before removing it from the system. This\ncan be set to \"infinity\" which sets it to the max value of 2147483647 sec‐\nonds, or 68 years. It will be automatically adjusted to the overall retry\ninterval nopathretry * pollinginterval if a number of retries is given\nwith nopathretry and the overall retry interval is longer than the speci‐\nfied devlosstmo value. The Linux kernel will cap this value to 600 if\nfastiofailtmo is not set. See KNOWN ISSUES.\n\nThe default is: 600\n\nehdeadline Specify the maximum number of seconds the SCSI layer will spend doing error\nhandling when scsi devices fail. After this timeout the scsi layer will per‐\nform a full HBA reset. Setting this may be necessary in cases where the\nrport is never lost, so fastiofailtmo and devlosstmo will never trig‐\nger, but (frequently do to load) scsi commands still hang. Note: when the\nscsi error handler performs the HBA reset, all target paths on that HBA will\nbe affected. ehdeadline should only be set in cases where all targets on\nthe affected HBAs are multipathed.\n\nThe default is: \n\nbindingsfile This option is deprecated, and will be removed in a future release. The\nfull pathname of the binding file to be used when the userfriendlynames\noption is set.\n\nThe default is: /etc/multipath/bindings\n\nwwidsfile This option is deprecated, and will be removed in a future release. The\nfull pathname of the WWIDs file, which is used by multipath to keep track of\nthe WWIDs for LUNs it has created multipath devices on in the past.\n\nThe default is: /etc/multipath/wwids\n\nprkeysfile This option is deprecated, and will be removed in a future release. The\nfull pathname of the prkeys file, which is used by multipathd to keep track\nof the persistent reservation key used for a specific WWID, when reserva‐\ntionkey is set to file.\n\nThe default is: /etc/multipath/prkeys\n\nlogcheckererr If set to once , multipathd logs the first path checker error at logging\nlevel 2. Any later errors are logged at level 3 until the device is re‐\nstored. If set to always , multipathd always logs the path checker error at\nlogging level 2.\n\nThe default is: always\n\nreservationkey This is the service action reservation key used by mpathpersist. It must be\nset for all multipath devices using persistent reservations, and it must be\nthe same as the RESERVATION KEY field of the PERSISTENT RESERVE OUT parame‐\nter list which contains an 8-byte value provided by the application client\nto the device server to identify the IT nexus. If the --param-aptpl option\nis used when registering the key with mpathpersist, :aptpl must be appended\nto the end of the reservation key.\n\nAlternatively, this can be set to file, which will store the RESERVATION KEY\nregistered by mpathpersist in the prkeysfile. multipathd will then use this\nkey to register additional paths as they appear. When the registration is\nremoved, the RESERVATION KEY is removed from the prkeysfile. The prkeys\nfile will automatically keep track of whether the key was registered with\n--param-aptpl.\n\nThe default is: \n\nalltgpt Set the 'all targets ports' flag when registering keys with mpathpersist.\nSome arrays automatically set and clear registration keys on all target\nports from a host, instead of per target port per host. The ALLTGPT flag\nmust be set to successfully use mpathpersist on these arrays. Setting this\noption is identical to calling mpathpersist with --param-alltgpt\n\nThe default is: no\n\nretainattachedhwhandler\n(Obsolete for kernels >= 4.3) If set to yes and the SCSI layer has already\nattached a hardwarehandler to the device, multipath will not force the de‐\nvice to use the hardwarehandler specified by mutipath.conf. If the SCSI\nlayer has not attached a hardware handler, multipath will continue to use\nits configured hardware handler.\n\nThe default is: yes\n\nImportant Note: Linux kernel 4.3 or newer always behaves as if \"retainat‐‐\ntachedhwhandler yes\" was set.\n\ndetectprio If set to yes , multipath will try to detect if the device supports SCSI-3\nALUA. If so, the device will automatically use the sysfs prioritizer if the\nrequired sysf attributes accessstate and preferredpath are supported, or\nthe alua prioritizer if not. If set to no , the prioritizer will be selected\nas usual.\n\nThe default is: yes\n\ndetectchecker if set to yes , multipath will try to detect if the device supports SCSI-3\nALUA. If so, the device will automatically use the tur checker. If set to no\n, the checker will be selected as usual.\n\nThe default is: yes\n\nforcesync If set to yes , multipathd will call the path checkers in sync mode only.\nThis means that only one checker will run at a time. This is useful in the\ncase where many multipathd checkers running in parallel causes significant\nCPU pressure.\n\nThe default is: no\n\nstricttiming If set to yes , multipathd will start a new path checker loop after exactly\none second, so that each path check will occur at exactly pollinginterval\nseconds. On busy systems path checks might take longer than one second; here\nthe missing ticks will be accounted for on the next round. A warning will\nbe printed if path checks take longer than pollinginterval seconds.\n\nThe default is: no\n\ndeferredremove If set to yes , multipathd will do a deferred remove instead of a regular\nremove when the last path device has been deleted. This means that if the\nmultipath device is still in use, it will be freed when the last user closes\nit. If path is added to the multipath device before the last user closes\nit, the deferred remove will be canceled.\n\nThe default is: no\n\npartitiondelimiter\nThis parameter controls how multipath chooses the names of partition devices\nof multipath maps if a multipath map is renamed (e.g. if a map alias is\nadded or changed). If this parameter is set to a string other than \"/UNSET/\"\n(even the empty string), multipath inserts that string between device name\nand partition number to construct the partition device name. Otherwise\n(i.e. if this parameter is unset or has the value \"/UNSET/\"), the behavior\ndepends on the map name: if it ends in a digit, a \"p\" is inserted between\nname and partition number; otherwise, the partition number is simply ap‐\npended. Distributions may use a non-null default value for this option; in\nthis case, the user must set it to \"/UNSET/\" to obtain the original \nbehavior. Use multipath -T to check the current settings.\n\nThe default is: \n\nconfigdir This option is deprecated, and will be removed in a future release. If set\nto anything other than \"\", multipath will search this directory alphabeti‐\ncally for file ending in \".conf\" and it will read configuration information\nfrom them, just as if it was in /etc/multipath.conf. configdir must either\nbe \"\" or a fully qualified directory name.\n\nThe default is: /etc/multipath/conf.d/\n\nsanpatherrthreshold\nIf set to a value greater than 0, multipathd will watch paths and check how\nmany times a path has been failed due to errors.If the number of failures on\na particular path is greater then the sanpatherrthreshold, then the path\nwill not reinstate till sanpatherrrecoverytime. These path failures\nshould occur within a sanpatherrforgetrate checks, if not we will con‐\nsider the path is good enough to reinstantate. See \"Shaky paths detection\"\nbelow.\n\nThe default is: no\n\nsanpatherrforgetrate\nIf set to a value greater than 0, multipathd will check whether the path\nfailures has exceeded the sanpatherrthreshold within this many checks\ni.e sanpatherrforgetrate . If so we will not reinstante the path till\nsanpatherrrecoverytime. See \"Shaky paths detection\" below.\n\nThe default is: no\n\nsanpatherrrecoverytime\nIf set to a value greater than 0, multipathd will make sure that when path\nfailures has exceeded the sanpatherrthreshold within sanpatherrfor‐\ngetrate then the path will be placed in failed state for sanpatherrre‐\ncoverytime duration.Once sanpatherrrecoverytime has timeout we will\nreinstante the failed path . sanpatherrrecoverytime value should be in\nsecs. See \"Shaky paths detection\" below.\n\nThe default is: no\n\nmarginalpathdoublefailedtime\nOne of the four parameters of supporting path check based on accounting IO\nerror such as intermittent error. When a path failed event occurs twice in\nmarginalpathdoublefailedtime seconds due to an IO error and all the\nother three parameters are set, multipathd will fail the path and enqueue\nthis path into a queue of which members are sent a couple of continuous di‐\nrect reading asynchronous IOs at a fixed sample rate of 10HZ to start IO er‐\nror accounting process. See \"Shaky paths detection\" below.\n\nThe default is: no\n\nmarginalpatherrsampletime\nOne of the four parameters of supporting path check based on accounting IO\nerror such as intermittent error. If it is set to a value no less than 120,\nwhen a path fail event occurs twice in marginalpathdoublefailedtime sec‐\nond due to an IO error, multipathd will fail the path and enqueue this path\ninto a queue of which members are sent a couple of continuous direct reading\nasynchronous IOs at a fixed sample rate of 10HZ to start the IO accounting\nprocess for the path will last for marginalpatherrsampletime. If the\nrate of IO error on a particular path is greater than the mar‐\nginalpatherrratethreshold, then the path will not reinstate for mar‐\nginalpatherrrecheckgaptime seconds unless there is only one active\npath. After marginalpatherrrecheckgaptime expires, the path will be re‐\nqueueed for rechecking. If checking result is good enough, the path will be\nreinstated. See \"Shaky paths detection\" below.\n\nThe default is: no\n\nmarginalpatherrratethreshold\nThe error rate threshold as a permillage (1/1000). One of the four parame‐\nters of supporting path check based on accounting IO error such as intermit‐\ntent error. Refer to marginalpatherrsampletime. If the rate of IO errors\non a particular path is greater than this parameter, then the path will not\nreinstate for marginalpatherrrecheckgaptime seconds unless there is\nonly one active path. See \"Shaky paths detection\" below.\n\nThe default is: no\n\nmarginalpatherrrecheckgaptime\nOne of the four parameters of supporting path check based on accounting IO\nerror such as intermittent error. Refer to marginalpatherrsampletime. If\nthis parameter is set to a positive value, the failed path of which the IO\nerror rate is larger than marginalpatherrratethreshold will be kept in\nfailed state for marginalpatherrrecheckgaptime seconds. When mar‐\nginalpatherrrecheckgaptime seconds expires, the path will be requeueed\nfor checking. If checking result is good enough, the path will be rein‐\nstated, or else it will keep failed. See \"Shaky paths detection\" below.\n\nThe default is: no\n\ndelaywatchchecks\nThis option is deprecated, and mapped to sanpatherrforgetrate. If this\nis set to a value greater than 0 and no sanpatherr options are set,\nsanpatherrforgetrate will be set to the value of delaywatchchecks and\nsanpatherrthreshold will be set to 1. See the sanpatherrforgetrate\nand sanpatherrthreshold options, and \"Shaky paths detection\" below for\nmore information.\n\nThe default is: no\n\ndelaywaitchecks\nThis option is deprecated, and mapped to sanpatherrrecoverytime. If\nthis is set to a value greater than 0 and no sanpatherr options are set,\nsanpatherrrecoverytime will be set to the value of delaywaitchecks\ntimes maxpollinginterval. This will give approximately the same wait time\nas delaywaitchecks previously did. Also, sanpatherrthreshold will be\nset to 1. See the sanpatherrrecoverytime and sanpatherrthreshold op‐\ntions, and \"Shaky paths detection\" below for more information.\n\nThe default is: no\n\nmarginalpathgroups\nIf set to no, the delay*checks, marginalpath*, and sanpatherr* op‐\ntions will keep marginal, or \"shaky\", paths from being reinstated until they\nhave been monitored for some time. This can cause situations where all non-\nmarginal paths are down, and no paths are usable until multipathd detects\nthis and reinstates a marginal path. If the multipath device is not config‐\nured to queue IO in this case, it can cause IO errors to occur, even though\nthere are marginal paths available. However, if this option is set to yes,\nwhen one of the marginal path detecting methods determines that a path is\nmarginal, it will be reinstated and placed in a seperate pathgroup that will\nonly be used after all the non-marginal pathgroups have been tried first.\nThis prevents the possibility of IO errors occuring while marginal paths are\nstill usable. After the path has been monitored for the configured time, and\nis declared healthy, it will be returned to its normal pathgroup. See \"Shaky\npaths detection\" below for more information.\n\nThe default is: no\n\nfindmultipaths This option controls whether multipath and multipathd try to create multi‐\npath maps over non-blacklisted devices they encounter. This matters a) when\na device is encountered by multipath -u during udev rule processing (a de‐\nvice is blocked from further processing by higher layers - such as LVM - if\nand only if it´s considered a valid multipath device path), and b) when mul‐\ntipathd detects a new device. The following values are possible:\n\nstrict Both multipath and multipathd treat only such devices as multipath\ndevices which have been part of a multipath map previously, and\nwhich are therefore listed in the wwidsfile. Users can manually\nset up multipath maps using the multipathd add map command. Once\nset up manually, the map is remembered in the wwids file and will\nbe set up automatically in the future.\n\nno Multipath behaves like strict. Multipathd behaves like greedy.\n\nyes Both multipathd and multipath treat a device as multipath device\nif the conditions for strict are met, or if at least two non-\nblacklisted paths with the same WWID have been detected.\n\ngreedy Both multipathd and multipath treat every non-blacklisted device\nas multipath device path.\n\nsmart This differs from findmultipaths yes only in the way it treats\nnew devices for which only one path has been detected yet. When\nsuch a device is first encounted in udev rules, it is treated as a\nmultipath device. multipathd waits whether additional paths with\nthe same WWID appears. If that happens, it sets up a multipath\nmap. If it doesn´t happen until a timeout expires, or if setting\nup the map fails, a new uevent is triggered for the device; at\nsecond encounter in the udev rules, the device will be treated as\nnon-multipath and passed on to upper layers. Note: this may cause\ndelays during device detection if there are single-path devices\nwhich aren´t blacklisted.\n\nThe default is: strict\n\nfindmultipathstimeout\nTimeout, in seconds, to wait for additional paths after detecting the first\none, if findmultipaths \"smart\" (see above) is set. If the value is posi‐‐\ntive, this timeout is used for all unknown, non-blacklisted devices encoun‐\ntered. If the value is negative (recommended), it's only applied to \"known\"\ndevices that have an entry in multipath's hardware table, either in the\nbuilt-in table or in a device section; other (\"unknown\") devices will use a\ntimeout of only 1 second to avoid booting delays. The value 0 means \"use the\nbuilt-in default\". If findmultipath has a value other than smart, this op‐\ntion has no effect.\n\nThe default is: -10 (10s for known and 1s for unknown hardware)\n\nuxsocktimeout CLI receive timeout in milliseconds. For larger systems CLI commands might\ntimeout before the multipathd lock is released and the CLI command can be\nprocessed. This will result in errors like \"timeout receiving packet\" to be\nreturned from CLI commands. In these cases it is recommended to increase\nthe CLI timeout to avoid those issues.\n\nThe default is: 4000\n\nretriggertries Sets the number of times multipathd will try to retrigger a uevent to get\nthe WWID.\n\nThe default is: 3\n\nretriggerdelay Sets the amount of time, in seconds, to wait between retriggers.\n\nThe default is: 10\n\nmissinguevwaittimeout\nControls how many seconds multipathd will wait, after a new multipath device\nis created, to receive a change event from udev for the device, before auto‐\nmatically enabling device reloads. Usually multipathd will delay reloads on\na device until it receives a change uevent from the initial table load.\n\nThe default is: 30\n\nskipkpartx If set to yes , kpartx will not automatically create partitions on the de‐\nvice.\n\nThe default is: no\n\ndisablechangedwwids\nThis option is deprecated and ignored. If the WWID of a path suddenly\nchanges, multipathd handles it as if it was removed and then added again.\n\nremoveretries This sets how may times multipath will retry removing a device that is in-\nuse. Between each attempt, multipath will sleep 1 second.\n\nThe default is: 0\n\nmaxsectorskb Sets the maxsectorskb device parameter on all path devices and the multi‐\npath device to the specified value.\n\nThe default is: in /sys/block//queue/maxsectorskb\n\nghostdelay Sets the number of seconds that multipath will wait after creating a device\nwith only ghost paths before marking it ready for use in systemd. This gives\nthe active paths time to appear before the multipath runs the hardware han‐\ndler to switch the ghost paths to active ones. Setting this to 0 or no makes\nmultipath immediately mark a device with only ghost paths as ready.\n\nThe default is: no\n\nenableforeign Enables or disables foreign libraries (see section FOREIGN MULTIPATH SUPPORT\nbelow). The value is a regular expression; foreign libraries are loaded if\ntheir name (e.g. \"nvme\") matches the expression. By default, no foreign li‐\nbraries are enabled. Set this to \"nvme\" to enable NVMe native multipath sup‐\nport, or \".*\" to enable all foreign libraries.\n\nThe default is: \"NONE\"\n\nrecheckwwid If set to yes, when a failed path is restored, its wwid is rechecked. If the\nwwid has changed, the path is removed from the current multipath device, and\nre-added as a new path. Multipathd will also recheck a path's wwid if it is\nmanually re-added. This option only works for SCSI devices that are config‐\nured to use the default uidattribute, IDSERIAL, or sysfs for getting their\nwwid.\n\nThe default is: no\n\n\nblacklist and blacklistexceptions sections\nThe blacklist section is used to exclude specific devices from the multipath topology. It is\nmost commonly used to exclude local disks or non-disk devices (such as LUNs for the storage\narray controller) from being handled by multipath-tools.\n\nIn the blacklist and blacklistexceptions sections, starting a quoted value with an exclama‐\ntion mark \"!\" will invert the matching of the rest of the regular expression. For instance,\n\"!^sd[a-z]\" will match all values that do not start with \"sd[a-z]\". The exclamation mark can\nbe escaped \"\\!\" to match a literal ! at the start of a regular expression. Note: The exclama‐\ntion mark must be inside quotes, otherwise it will be treated as starting a comment.\n\nThe blacklistexceptions section is used to revert the actions of the blacklist section. This\nallows one to selectively include (\"whitelist\") devices which would normally be excluded via\nthe blacklist section. A common usage is to blacklist \"everything\" using a catch-all regular\nexpression, and create specific blacklistexceptions entries for those devices that should be\nhandled by multipath-tools.\n\nThe following keywords are recognized in both sections. The defaults are empty unless explic‐\nitly stated.\n\ndevnode Regular expression matching the device nodes to be excluded/included.\n\nThe default blacklist consists of the regular expression \"!^(sd[a-z]|dasd[a-\nz]|nvme[0-9])\". This causes all device types other than scsi, dasd, and nvme\nto be excluded from multipath handling by default.\n\nwwid Regular expression for the World Wide Identifier of a device to be ex‐\ncluded/included.\n\ndevice Subsection for the device description. This subsection recognizes the vendor\nand product keywords. Both are regular expressions. For a full description\nof these keywords please see the devices section description.\n\nproperty Regular expression for an udev property. All devices that have matching udev\nproperties will be excluded/included. The handling of the property keyword\nis special, because devices must have at least one whitelisted udev prop‐\nerty; otherwise they're treated as blacklisted, and the message \"black‐\nlisted, udev property missing\" is displayed in the logs.\n\nNote: The behavior of this option has changed in multipath-tools 0.8.2 com‐\npared to previous versions. Blacklisting by missing properties is only ap‐\nplied to devices which do have the property specified by uidattribute (e.g.\nIDSERIAL) set. Previously, it was applied to every device, possibly causing\ndevices to be blacklisted because of temporary I/O error conditions.\n\nThe default blacklist exception is: (SCSIIDENT|IDWWN), causing well-be‐\nhaved SCSI devices and devices that provide a WWN (World Wide Number) to be\nincluded, and all others to be excluded.\n\nprotocol Regular expression for the protocol of a device to be excluded/included.\n\nThe protocol strings that multipath recognizes are scsi:fcp, scsi:spi,\nscsi:ssa, scsi:sbp, scsi:srp, scsi:iscsi, scsi:sas, scsi:adt, scsi:ata,\nscsi:unspec, ccw, cciss, nvme, and undef. The protocol that a path is using\ncan be viewed by running multipathd show paths format \"%d %P\"\n\nFor every device, these 5 blacklist criteria are evaluated in the the order \"property, dev‐\nnode, device, protocol, wwid\". If a device turns out to be blacklisted by any criterion, it's\nexcluded from handling by multipathd, and the later criteria aren't evaluated any more. For\neach criterion, the whitelist takes precedence over the blacklist if a device matches both.\n\nNote: Besides the blacklist and whitelist, other configuration options such as findmulti‐\npaths have an impact on whether or not a given device is handled by multipath-tools.\n" }, { "name": "multipaths section", "content": "The multipaths section allows setting attributes of multipath maps. The attributes that are\nset via the multipaths section (see list below) take precedence over all other configuration\nsettings, including those from the overrides section.\n\nThe only recognized attribute for the multipaths section is the multipath subsection. If\nthere are multiple multipath subsections matching a given WWID, the contents of these sec‐\ntions are merged, and settings from later entries take precedence.\n\nThe multipath subsection recognizes the following attributes:\n\nwwid (Mandatory) World Wide Identifier. Detected multipath maps are matched\nagains this attribute. Note that, unlike the wwid attribute in the black‐\nlist section, this is not a regular expression or a substring; WWIDs must\nmatch exactly inside the multipaths section.\n\nalias Symbolic name for the multipath map. This takes precedence over a an entry\nfor the same WWID in the bindingsfile.\n\nThe following attributes are optional; if not set the default values are taken from the over‐\nrides, devices, or defaults section:\n\npathgroupingpolicy\npathselector\nprio\nprioargs\nfailback\nrrweight\nnopathretry\nrrminio\nrrminiorq\nflushonlastdel\nfeatures\nreservationkey\nuserfriendlynames\ndeferredremove\nsanpatherrthreshold\nsanpatherrforgetrate\nsanpatherrrecoverytime\nmarginalpatherrsampletime\nmarginalpatherrratethreshold\nmarginalpatherrrecheckgaptime\nmarginalpathdoublefailedtime\ndelaywatchchecks\ndelaywaitchecks\nskipkpartx\nmaxsectorskb\nghostdelay\n" }, { "name": "devices section", "content": "multipath-tools have a built-in device table with reasonable defaults for more than 100 known\nmultipath-capable storage devices. The devices section can be used to override these set‐\ntings. If there are multiple matches for a given device, the attributes of all matching en‐\ntries are applied to it. If an attribute is specified in several matching device subsec‐\ntions, later entries take precedence. Thus, entries in files under configdir (in reverse al‐\nphabetical order) have the highest precedence, followed by entries in multipath.conf; the\nbuilt-in hardware table has the lowest precedence. Inside a configuration file, later entries\nhave higher precedence than earlier ones.\n\nThe only recognized attribute for the devices section is the device subsection. Devices de‐\ntected in the system are matched against the device entries using the vendor, product, and\nrevision fields, which are all POSIX Extended regular expressions (see regex(7)).\n\nThe vendor, product, and revision fields that multipath or multipathd detect for devices in a\nsystem depend on the device type. For SCSI devices, they correspond to the respective fields\nof the SCSI INQUIRY page. In general, the command 'multipathd show paths format \"%d %s\"' com‐\nmand can be used to see the detected properties for all devices in the system.\n\nThe device subsection recognizes the following attributes:\n\nvendor (Mandatory) Regular expression to match the vendor name.\n\nproduct (Mandatory) Regular expression to match the product name.\n\nrevision Regular expression to match the product revision. If not specified, any re‐\nvision matches.\n\nproductblacklist\nProducts with the given vendor matching this string are blacklisted. This is\nequivalent to a device entry in the blacklist section with the vendor attri‐\nbute set to this entry's vendor, and the product attribute set to the value\nof productblacklist.\n\naliasprefix The userfriendlynames prefix to use for this device type, instead of the\ndefault \"mpath\".\n\nvpdvendor The vendor specific vpd page information, using the vpd page abbreviation.\nThe vpd page abbreviation can be found by running sgvpd -e. multipathd will\nuse this information to gather device specific information that can be dis‐\nplayed with the %g wilcard for the multipathd show maps format and multi‐\npathd show paths format commands. Currently only the hp3par vpd page is sup‐\nported.\n\nhardwarehandler The hardware handler to use for this device type. The following hardware\nhandler are implemented:\n\n1 emc (Hardware-dependent) Hardware handler for DGC class arrays as\nCLARiiON CX/AX and EMC VNX and Unity families.\n\n1 rdac (Hardware-dependent) Hardware handler for LSI/Engenio/NetApp\nRDAC class as NetApp SANtricity E/EF Series, and OEM arrays from\nIBM DELL SGI STK and SUN.\n\n1 hpsw (Hardware-dependent) Hardware handler for HP/COMPAQ/DEC HSG80\nand MSA/HSV arrays with Active/Standby mode exclusively.\n\n1 alua (Hardware-dependent) Hardware handler for SCSI-3 ALUA compatible\narrays.\n\n1 ana (Hardware-dependent) Hardware handler for NVMe ANA compatible\narrays.\n\nThe default is: \n\nImportant Note: Linux kernels 4.3 and newer automatically attach a device\nhandler to known devices (which includes all devices supporting SCSI-3 ALUA)\nand disallow changing the handler afterwards. Setting hardwarehandler for\nsuch devices on these kernels has no effect.\n\nThe following attributes are optional; if not set the default values are taken from the de‐\nfaults section:\n\npathgroupingpolicy\nuidattribute\ngetuidcallout\npathselector\npathchecker\nprio\nprioargs\nfeatures\nfailback\nrrweight\nnopathretry\nrrminio\nrrminiorq\nfastiofailtmo\ndevlosstmo\nflushonlastdel\nuserfriendlynames\nretainattachedhwhandler\ndetectprio\ndetectchecker\ndeferredremove\nsanpatherrthreshold\nsanpatherrforgetrate\nsanpatherrrecoverytime\nmarginalpatherrsampletime\nmarginalpatherrratethreshold\nmarginalpatherrrecheckgaptime\nmarginalpathdoublefailedtime\ndelaywatchchecks\ndelaywaitchecks\nskipkpartx\nmaxsectorskb\nghostdelay\nalltgpt\n" }, { "name": "overrides section", "content": "The overrides section recognizes the following optional attributes; if not set the values are\ntaken from the devices or defaults sections:\n\npathgroupingpolicy\nuidattribute\ngetuidcallout\npathselector\npathchecker\naliasprefix\nfeatures\nprio\nprioargs\nfailback\nrrweight\nnopathretry\nrrminio\nrrminiorq\nflushonlastdel\nfastiofailtmo\ndevlosstmo\nuserfriendlynames\nretainattachedhwhandler\ndetectprio\ndetectchecker\ndeferredremove\nsanpatherrthreshold\nsanpatherrforgetrate\nsanpatherrrecoverytime\nmarginalpatherrsampletime\nmarginalpatherrratethreshold\nmarginalpatherrrecheckgaptime\nmarginalpathdoublefailedtime\ndelaywatchchecks\ndelaywaitchecks\nskipkpartx\nmaxsectorskb\nghostdelay\nalltgpt\n" }, { "name": "WWID generation", "content": "Multipath uses a World Wide Identification (WWID) to determine which paths belong to the same\ndevice. Each path presenting the same WWID is assumed to point to the same device.\n\nThe WWID is generated by four methods (in the order of preference):\n\nuidattrs The WWID is derived from udev attributes by matching the device node name;\ncf uidattrs above.\n\ngetuidcallout Use the specified external program; cf getuidcallout above. Care should be\ntaken when using this method; the external program needs to be loaded from\ndisk for execution, which might lead to deadlock situations in an all-paths-\ndown scenario.\n\nuidattribute Use the value of the specified udev attribute; cf uidattribute above. This\nmethod is preferred to getuidcallout as multipath does not need to call any\nexternal programs here. However, under certain circumstances udev might not\nbe able to generate the requested variable.\n\nsysfs Try to determine the WWID from sysfs attributes. For SCSI devices, this\nmeans reading the Vital Product Data (VPD) page \"Device Identification\"\n(0x83).\n\nThe default settings (using udev and uidattribute configured from the built-in hardware ta‐\nble) should work fine in most scenarios. Users who want to enable uevent merging must set\nuidattrs.\n" }, { "name": "Shaky paths detection", "content": "A common problem in SAN setups is the occurence of intermittent errors: a path is unreach‐\nable, then reachable again for a short time, disappears again, and so forth. This happens\ntypically on unstable interconnects. It is undesirable to switch pathgroups unnecessarily on\nsuch frequent, unreliable events. multipathd supports three different methods for detecting\nthis situation and dealing with it. All methods share the same basic mode of operation: If a\npath is found to be \"shaky\" or \"flipping\", and appears to be in healthy status, it is not re‐\ninstated (put back to use) immediately. Instead, it is placed in the \"delayed\" state and\nwatched for some time, and only reinstated if the healthy state appears to be stable. If the\nmarginalpathgroups option is set, the path will reinstated immediately, but placed in a spe‐\ncial pathgroup for marginal paths. Marginal pathgroups will not be used until all other path‐\ngroups have been tried. At the time when the path would normally be reinstated, it will be\nreturned to its normal pathgroup. The logic of determining \"shaky\" condition, as well as the\nlogic when to reinstate, differs between the three methods.\n\n\"delaychecks\" failure tracking\nThis method is deprecated and mapped to the \"sanpatherr\" method. See the de‐\nlaywatchchecks and delaywaitchecks options above for more information.\n\n\n\"marginalpath\" failure tracking\nIf a second failure event (good->bad transition) occurs within marginalpathdou‐\nblefailedtime seconds after a failure, high-frequency monitoring is started for the\naffected path: I/O is sent at a rate of 10 per second. This is done for mar‐\nginalpatherrsampletime seconds. During this period, the path is not reinstated.\nIf the rate of errors remains below marginalpatherrratethreshold during the moni‐\ntoring period, the path is reinstated. Otherwise, it is kept in failed state for mar‐\nginalpatherrrecheckgaptime, and after that, it is monitored again. For this\nmethod, time intervals are measured in seconds.\n\n\"sanpatherr\" failure tracking\nmultipathd counts path failures for each path. Once the number of failures exceeds\nthe value given by sanpatherrthreshold, the path is not reinstated for\nsanpatherrrecoverytime seconds. While counting failures, multipathd \"forgets\" one\npast failure every \"sanpatherrforgetrate\" ticks; thus if errors don't occur more\noften then once in the forget rate interval, the failure count doesn't increase and\nthe threshold is never reached. Ticks are the time between path checks by multipathd,\nwhich is variable and controlled by the pollinginterval and maxpollinginterval pa‐\nrameters.\n\nThis method is deprecated in favor of the \"marginalpath\" failure tracking method,\nand only offered for backward compatibility.\n\nSee the documentation of the individual options above for details. It is strongly discour‐‐\naged to use more than one of these methods for any given multipath map, because the two con‐\ncurrent methods may interact in unpredictable ways. If the \"marginalpath\" method is active,\nthe \"sanpatherr\" parameters are implicitly set to 0.\n" } ] }, "FOREIGN MULTIPATH SUPPORT": { "content": "multipath and multipathd can load \"foreign\" libraries to add support for other multipathing\ntechnologies besides the Linux device mapper. Currently this support is limited to printing\ndetected information about multipath setup. In topology output, the names of foreign maps are\nprefixed by the foreign library name in square brackets, as in this example:\n\n# multipath -ll\nuuid.fedcba98-3579-4567-8765-123456789abc [nvme]:nvme4n9 NVMe,Some NVMe controller,FFFFFFFF\nsize=167772160 features='n/a' hwhandler='ANA' wp=rw\n|-+- policy='n/a' prio=50 status=optimized\n| `- 4:38:1 nvme4c38n1 0:0 n/a optimized live\n`-+- policy='n/a' prio=50 status=optimized\n`- 4:39:1 nvme4c39n1 0:0 n/a optimized live\n\nThe \"nvme\" foreign library provides support for NVMe native multipathing in the kernel. It is\npart of the standard multipath package.\n", "subsections": [] }, "KNOWN ISSUES": { "content": "The usage of queueifnopath option can lead to D state processes being hung and not kill‐\nable in situations where all the paths to the LUN go offline. It is advisable to use the\nnopathretry option instead.\n\nThe use of queueifnopath or nopathretry might lead to a deadlock if the devlosstmo\nsetting results in a device being removed while I/O is still queued. The multipath daemon\nwill update the devlosstmo setting accordingly to avoid this deadlock. Hence if both values\nare specified the order of precedence is nopathretry, queueifnopath, devlosstmo.\n", "subsections": [] }, "SEE ALSO": { "content": "udev(8), dmsetup(8), multipath(8), multipathd(8).\n", "subsections": [] }, "AUTHORS": { "content": "multipath-tools was developed by Christophe Varoqui, and\nothers.\n\n\n\nLinux 2021-09-08 MULTIPATH.CONF(5)", "subsections": [] } } } }