MDADM.CONF(5)

💖 mdadm.conf(5) 💖 — File Formats Manual

okay bestie, pull up a chair — this is the config file page from the File Formats Manual, and it's literally the whole personality of your software RAID setup. Let's get into it. 💅

---

🎀 NAME 🎀

mdadm.conf — configuration for management of Software RAID with mdadm. That's it, that's the whole vibe. 👑

---

✨ SYNOPSIS ✨

/etc/mdadm.conf

---

🦋 DESCRIPTION 🦋

mdadm is a tool for creating, managing, and monitoring RAID devices using the md driver in Linux.

And some of the common tasks — like assembling all arrays, which, yes, is as iconic as it sounds — get SO much simpler when you describe the devices and arrays in this configuration file. No cap. 🫶

🌸 SYNTAX 🌸

Think of the file as a whole collection of words separated by white space (space, tab, or newline).

• Any word that begins with a hash sign (#) starts a comment — that word, together with the remainder of the line, is ignored. bye. ✌️
• Spaces can be included in a word using quotation characters. Either single quotes (') or double quotes (") may be used.
• All the characters from one quotation character to the next identical character are protected, and will not be used to separate words or to start new quoted strings.
• To include a single quote it must be between double quotes.
• To include a double quote it must be between single quotes.
• Any line that starts with white space (space or tab) is treated as though it were a continuation of the previous line.
• Empty lines are ignored.
• But otherwise, each (non-continuation) line must start with a keyword as listed below.
• The keywords are case insensitive and can be abbreviated to 3 characters.

The keywords are:

🖥️ DEVICE

A device line lists the devices (whole devices or partitions) that might contain a component of an MD array. When looking for the components of an array, mdadm will scan these devices — or any devices listed on the command line.

The device line may contain a number of different devices (separated by spaces), and each device name can contain wild cards as defined by glob(7).

Also, there may be several device lines present in the file. Go off, honestly. ✨

Alternatively, a device line can contain either or both of the words containers and partitions.

• containers — this will cause mdadm to look for assembled CONTAINER arrays and include them as a source for assembling further arrays.
• partitions — this will cause mdadm to read /proc/partitions and include all devices and partitions found therein. mdadm does not use the names from /proc/partitions but only the major and minor device numbers. It scans /dev to find the name that matches the numbers.

🚨 If no DEVICE line is present in any config file, then DEVICE partitions containers is assumed.

For example:

DEVICE /dev/hda* /dev/hdc*
DEV    /dev/sd*
DEVICE /dev/disk/by-path/pci*
DEVICE partitions

🔥 ARRAY

The ARRAY lines identify actual arrays. The second word on the line may be the name of the device where the array is normally assembled, such as /dev/md1 or /dev/md/backup. If the name does not start with a slash (/), it is treated as being in /dev/md/. Alternately, the word <ignore> (complete with angle brackets) can be given, in which case any array which matches the rest of the line will never be automatically assembled. And if no device name is given, mdadm will use various heuristics to determine an appropriate name. 🤷‍♀️

Subsequent words identify the array, or identify the array as a member of a group. If multiple identities are given, then a component device must match ALL identities to be considered a match. Each identity word has a tag, an equals sign, and some value. The tags are:

• uuid= — the value should be a 128 bit uuid in hexadecimal, with punctuation interspersed if desired. This must match the uuid stored in the superblock.
• super-minor= — the value is an integer which indicates the minor number that was stored in the superblock when the array was created. When an array is created as /dev/mdX, then the minor number X is stored.
• devices= — the value is a comma separated list of device names or device name patterns. Only devices with names which match one entry in the list will be used to assemble the array. Note that the devices listed there must also be listed on a DEVICE line.
• level= — the value is a RAID level. This is not normally used to identify an array, but is supported so that the output of

mdadm --examine --scan

can be use directly in the configuration file.

• num-devices= — the value is the number of devices in a complete active array. As with level= this is mainly for compatibility with the output of

mdadm --examine --scan.

• spares= — the value is a number of spare devices to expect the array to have. The sole use of this keyword and value is as follows: mdadm --monitor will report an array if it is found to have fewer than this number of spares when --monitor starts or when --oneshot is used.
• spare-group= — the value is a textual name for a group of arrays. All arrays with the same spare-group name are considered to be part of the same group. The significance of a group of arrays is that mdadm will, when monitoring the arrays, move a spare drive from one array in a group to another array in that group if the first array had a failed or missing drive but no spare. the teamwork… we love to see it 🫶
• bitmap= — the option specifies a file in which a write-intent bitmap should be found. When assembling the array, mdadm will provide this file to the md driver as the bitmap file. This has the same function as the --bitmap-file option to --assemble.
• metadata= — specify the metadata format that the array has. This is mainly recognised for comparability with the output of mdadm -Es.
• container= — specify that this array is a member array of some container. The value given can be either a path name in /dev, or a UUID of the container array.
• member= — specify that this array is a member array of some container. Each type of container has some way to enumerate member arrays, often a simple sequence number. The value identifies which member of a container the array is. It will usually accompany a container= word.

📧 MAILADDR

The mailaddr line gives an E-mail address that alerts should be sent to when mdadm is running in --monitor mode (and was given the --scan option). There should only be one MAILADDR line and it should have only one address. Any subsequent addresses are silently ignored. 🤫

💌 MAILFROM

The mailfrom line (which can only be abbreviated to at least 5 characters) gives an address to appear in the "From" address for alert mails. This can be useful if you want to explicitly set a domain, as the default from address is "root" with no domain. All words on this line are catenated with spaces to form the address.

Note that this value cannot be set via the mdadm commandline. It is only settable via the config file. There should only be one MAILADDR line and it should have only one address. Any subsequent addresses are silently ignored.

🏃 PROGRAM

The program line gives the name of a program to be run when mdadm --monitor detects potentially interesting events on any of the arrays that it is monitoring. This program gets run with two or three arguments, they being the Event, the md device, and possibly the related component device.

There should only be one program line and it should be given only one program. Any subsequent programs are silently ignored.

✨ CREATE

The create line gives default values to be used when creating arrays, new members of arrays, and device entries for arrays.

There should only be one create line. Any subsequent lines will override the previous settings.

Keywords used in the CREATE line and supported values are:

• owner=, group= — these can give user/group ids or names to use instead of system defaults (root/wheel or root/disk).
• mode= — an octal file mode such as 0660 can be given to override the default of 0600.
• metadata= — the name of the metadata format to use if none is explicitly given. This can be useful to impose a system-wide default of version-1 superblocks.
• names=yes — it has been possible to create md devices with a name like md_home rather than just a number, like md3. mdadm will use the numeric alternative by default as other tools that interact with md arrays may expect only numbers. If names=yes is given in mdadm.conf then mdadm will use a name when appropriate. If names=no is given, then non-numeric md device names will not be used even if the default changes in a future release of mdadm.
• bbl=no — by default, mdadm will reserve space for a bad block list (bbl) on all devices included in or added to any array that supports them. Setting bbl=no will prevent this, so newly added devices will not have a bad block log.

🏠 HOMEHOST

The homehost line gives a default value for the --homehost= option to mdadm. There should normally be only one other word on the line. It should either be a host name, or one of the special words <system>, <none> and <ignore>.

• If <system> is given, then the gethostname(2) systemcall is used to get the host name. This is the default.
• If <ignore> is given, then a flag is set so that when arrays are being auto-assembled the checking of the recorded homehost is disabled. If <ignore> is given it is also possible to give an explicit name which will be used when creating arrays. This is the only case when there can be more that one other word on the HOMEHOST line. If there are other words, or other HOMEHOST lines, they are silently ignored.
• If <none> is given, then the default of using gethostname(2) is over-ridden and no homehost name is assumed.

---

When arrays get created, that host name gets baked right into the metadata, bestie. So later, when mdadm does auto-assembly, any array that doesn't have the correct homehost name recorded in its metadata gets assembled under a "foreign" name instead. And a "foreign" name always ends with a digit string, with an underscore in front of it, so it can never be confused with a possible local name. Like /dev/md/1_1 or /dev/md/home_0. cute little naming convention fr fr

💫 HOMECLUSTER 💫

The homecluster line hands mdadm a default value for the --homecluster= option. It's literally just telling it the cluster name for the md device. The md device can only be assembled on the cluster that matches the name you specified. If homcluster isn't provided, mdadm tries to sniff out the cluster name all on its own. ✨

There should only be one homecluster line. Any subsequent lines are silently ignored. Not even a warning. Just vibes. 🙈

🌈 AUTO 🌈

Okay so you give a list of metadata format names, and each one gets a plus or minus sign in front of it. You can also use the word homehost, and all is allowed too — both of them preceded by a plus or minus sign as well. all is usually last.

Here's the tea: when mdadm is auto-assembling an array (via --assemble or --incremental) and it finds metadata of a given type, it checks that metadata type against the ones you listed on this line. First match wins, and all matches literally anything. If the match had a plus sign in front of it → auto assembly is allowed. If it had a minus sign → auto assembly is not allowed, periodt. And if no match is found at all? Auto assembly is allowed.

If the metadata says the array was created for this host, and the word homehost shows up before any other match, then that array is treated as a totally valid candidate for auto-assembly. 👑

What's this even for? Glad you asked bestie:

• Turning off all auto-assembly, so only arrays explicitly listed in mdadm.conf or on the command line ever get assembled.
• Turning off assembly of certain metadata types that might be handled by other software.
• Turning off assembly of all foreign arrays — because normally those do get assembled, they just get a non-deterministic name in /dev/md/.

The known metadata types are 0.90, 1.x, ddf, imsm. That's the whole guest list.

AUTO should be given at most once. Subsequent lines are silently ignored. Which means a later config file in a config directory will not overwrite the setting from an earlier config file. Sorry girl, first one wins. 💅

👑 POLICY 👑

This one specifies what automatic behavior is allowed on devices that newly appear in the system, and it also gives you a way to mark spares that can be moved to other arrays, plus the migration domains. A domain gets defined through a policy line by specifying a domain name for a number of paths from /dev/disk/by-path/. A device can belong to several domains (she contains multitudes 💜). And the domain of an array is the union of the domains of all the devices in that array. A spare can be automatically moved from one array to another if the set of the destination array's domains contains all the domains of the new disk, or if both arrays have the same spare-group.

To update the hot plug configuration you gotta run mdadm --udev-rules=<path_to_file> — like /etc/udev/rules.d/65-md-bare.rules — after changing the config file. And also run udevadm control --reload, otherwise bestie you're rebooting. 😭

Keywords used in the POLICY line, and the values they support:

• domain= — literally any arbitrary string, go feral.
• metadata= — one of 0.9, 1.x, ddf or imsm.
• path= — a file glob matching anything from /dev/disk/by-path.
• type= — either disk or part.
• action= — include, re-add, spare, spare-same-slot, or force-spare.
• auto= — yes, no, or homehost.

The action item decides the automatic behavior allowed for devices matching the path and type on that same line. If a device matches several lines with different actions, then the most permissive one applies. And the ordering of policy lines is irrelevant to the end result — she's order-agnostic, we love that. ✨

include

Allows adding a disk to an array if the metadata on that disk matches that array.

re-add

Will include the device in the array if it appears to be a current member, or a member that was recently removed, and the array has a write-intent-bitmap to allow the re-add functionality.

spare

As above, and additionally: if the device is bare it can become a spare if there's any array that it's a candidate for, based on domains and metadata.

spare-same-slot

As above, and additionally: if the given slot was used by an array that went degraded recently and the device you plugged in has no metadata, then it will automatically be added to that array (or its container). 🫶

force-spare

As above, and the disk will become a spare in the remaining cases.

🧚 PART-POLICY 🧚

This is basically POLICY's twin — it's similar and accepts the same keyword assignments. It lets you apply a consistent set of policies to each of the partitions of a device.

A PART-POLICY line should set type=disk and identify the path to one or more disk devices. Every partition on those disks gets treated according to the action= setting from this line. And if a domain is set in the line, then the domain associated with each partition is based on that domain — but with "-partN" appended, where N is the partition number of the partition that was found. 🎀

🌸 SYSFS 🌸

The SYSFS line lists custom values of an MD device's sysfs attributes, which get stored in sysfs after the array is assembled. Multiple lines are allowed, and every single line has to contain the uuid or the name of the device it relates to. Lines are applied in reverse order.

• uuid= — hexadecimal identifier of the MD device. This has to match the uuid stored in the superblock.
• name= — name of the MD device as it was given to mdadm when the array was created. It will be ignored if uuid is not empty.

⏰ MONITORDELAY ⏰

The monitordelay line gives a delay in seconds that mdadm shall wait before pooling md arrays, when mdadm is running in --monitor mode. The -d/--delay command line argument takes precedence over the config file. So the CLI is the main character here, always. 💅

If multiple MINITORDELAY lines are provided, only the first non-zero value is considered.

yes bestie, it's spelled MINITORDELAY in the source. we do not make the rules. 💀

🔒 ENCRYPTION_NO_VERIFY 🔒

ENCRYPTION_NO_VERIFY disables encryption verification for devices with particular encryption support detected. Currently, only verification of SATA OPAL encryption can be disabled. It does not disable ATA security encryption verification. And currently it's only effective for IMSM metadata. The available parameter is sata_opal.

🔍 PROBING 🔍

The PROBING line gives you options to configure device probing.

ddf_extended

Use the extended algorithm to detect DDF headers on disks. Instead of looking for the DDF super block only in the last block of the device, it scans the last 32 MB. This lets you detect metadata created by some RAID controllers — at the cost of slower probing. (Trade-offs, bestie. It's giving "worth it" or it's giving "why is this taking so long" and only you can decide. 😭)

---

📁 FILES 📁

/etc/mdadm.conf

The default config file location, used when mdadm is running without the --config option.

/etc/mdadm.conf.d

The default directory with config files. Used when mdadm is running without the --config option, after successfully reading the /etc/mdadm.conf default config file. Files in that directory are read in lexical order.

/etc/mdadm/mdadm.conf

Alternative config file that is read when mdadm is running without the --config option and the /etc/mdadm.conf default config file was not opened successfully.

/etc/mdadm/mdadm.conf.d

The alternative directory with config files. Used when mdadm is running without the --config option, after reading the /etc/mdadm/mdadm.conf alternative config file — whether that reading was successful or not. Files in that directory are read in lexical order.

---

🌟 EXAMPLE 🌟

DEVICE /dev/sd[bcdjkl]1
DEVICE /dev/hda1 /dev/hdb1

# /dev/md0 is known by its UUID.
ARRAY /dev/md0 UUID=3aaa0122:29827cfa:5331ad66:ca767371
# /dev/md1 contains all devices with a minor number of
#   1 in the superblock.
ARRAY /dev/md1 superminor=1
# /dev/md2 is made from precisely these two devices
ARRAY /dev/md2 devices=/dev/hda1,/dev/hdb1

# /dev/md4 and /dev/md5 are a spare-group and spares
#  can be moved between them
ARRAY /dev/md4 uuid=b23f3c6d:aec43a9f:fd65db85:369432df
           spare-group=group1
ARRAY /dev/md5 uuid=19464854:03f71b1b:e0df2edd:246cc977
           spare-group=group1
# /dev/md/home is created if need to be a partitionable md array
# any spare device number is allocated.
ARRAY /dev/md/home UUID=9187a482:5dde19d9:eea3cc4a:d646ab8b
           auto=part
# One domain comprising of devices attached to specified paths is defined.
# Bare device matching first path will be made an imsm spare on hot plug.
# If more than one array is created on devices belonging to domain1 and
# one of them becomes degraded, then any imsm spare matching any path for
# given domain name can be migrated.
POLICY domain=domain1 metadata=imsm path=pci-0000:00:1f.2-scsi-*
           action=spare
POLICY domain=domain1 metadata=imsm path=pci-0000:04:00.0-scsi-[01]*
           action=include
MAILADDR root@mydomain.tld
PROGRAM /usr/sbin/handle-mdadm-events
CREATE group=system mode=0640 auto=part-8
HOMEHOST <system>
AUTO +1.x homehost -all
SYSFS name=/dev/md/raid5 group_thread_cnt=4 sync_speed_max=1000000
SYSFS uuid=bead5eb6:31c17a27:da120ba2:7dfda40d group_thread_cnt=4
sync_speed_max=1000000
MONITORDELAY 60
ENCRYPTION_NO_VERIFY sata_opal

---

🦋 SEE ALSO 🦋

mdadm(8), md(4).

---

and that's the whole thing, bestie!! you read the entire MDADM.CONF(5) page and you understood every single line of it, I'm so proud of you it's unreal 🥹💖 go forth and configure your arrays, you absolute icon 👑✨ 💅