Frequently Asked Questions - Revision history

Spitzcor: /* What do all these acronyms mean? */

2024-06-27T19:06:40Z

What do all these acronyms mean?

← Older revision		Revision as of 12:06, 27 June 2024
Line 132:		Line 132:
	* '''OST''' - '''O'''bject '''S'''torage '''T'''arget - storage device that holds the filesystem data (regular data files, not directories, xattrs, or other metadata)		* '''OST''' - '''O'''bject '''S'''torage '''T'''arget - storage device that holds the filesystem data (regular data files, not directories, xattrs, or other metadata)
	* '''PTLRPC''' - '''P'''or'''t'''a'''l''' '''RPC''' - remote procedure call protocol layered over LNet		* '''PTLRPC''' - '''P'''or'''t'''a'''l''' '''RPC''' - remote procedure call protocol layered over LNet
	* '''QMT''' - '''Q'''uota '''M'''aster '''T'''~~target~~ - Global quota limits are stored in dedicated index files (there is one such index per quota type) on the quota master target (aka QMT)		* '''QMT''' - '''Q'''uota '''M'''aster '''T'''arget - Global quota limits are stored in dedicated index files (there is one such index per quota type) on the quota master target (aka QMT)
	* '''SEQ''' - '''SEQ'''uence number of FIDs that are allocated to a client or server for its own exclusive use. The SEQ controller on MDT0000 manages the entire 64-bit SEQ space, and grants large sub-ranges to MDTs and OSTs, which may in turn grant sub-sub-ranges to individual clients.		* '''SEQ''' - '''SEQ'''uence number of FIDs that are allocated to a client or server for its own exclusive use. The SEQ controller on MDT0000 manages the entire 64-bit SEQ space, and grants large sub-ranges to MDTs and OSTs, which may in turn grant sub-sub-ranges to individual clients.

Hornc: /* What do all these acronyms mean? */

2024-06-27T19:06:19Z

What do all these acronyms mean?

← Older revision		Revision as of 12:06, 27 June 2024
Line 115:		Line 115:
	* '''LOD''' - '''L'''ogical '''O'''bject '''D'''evice - MDS software layer that handles access to multiple MDTs and multiple OSTs		* '''LOD''' - '''L'''ogical '''O'''bject '''D'''evice - MDS software layer that handles access to multiple MDTs and multiple OSTs
	* '''LOV''' - '''L'''ogical '''O'''bject '''V'''olume - client software layer that handles client (llite) access to multiple OSTs		* '''LOV''' - '''L'''ogical '''O'''bject '''V'''olume - client software layer that handles client (llite) access to multiple OSTs
	* '''LWP''' - '''L'''ight '''W'''eight '''P'''~~rotocol~~ - a 'lightweight' client co-located on servers that cannot perform replay or recovery		* '''LWP''' - '''L'''ight '''W'''eight '''P'''roxy - a 'lightweight' client co-located on servers that cannot perform replay or recovery
	* '''MDC''' - '''M'''eta'''D'''ata '''C'''lient - client software layer that interfaces to the MDS		* '''MDC''' - '''M'''eta'''D'''ata '''C'''lient - client software layer that interfaces to the MDS
	* '''MDD''' - '''M'''etadata '''D'''evice '''D'''river - MDS software layer that understands POSIX semantics for file access		* '''MDD''' - '''M'''etadata '''D'''evice '''D'''river - MDS software layer that understands POSIX semantics for file access

Spitzcor: Add LWP and QMT

2024-06-27T18:59:46Z

Add LWP and QMT

← Older revision		Revision as of 11:59, 27 June 2024
Line 115:		Line 115:
	* '''LOD''' - '''L'''ogical '''O'''bject '''D'''evice - MDS software layer that handles access to multiple MDTs and multiple OSTs		* '''LOD''' - '''L'''ogical '''O'''bject '''D'''evice - MDS software layer that handles access to multiple MDTs and multiple OSTs
	* '''LOV''' - '''L'''ogical '''O'''bject '''V'''olume - client software layer that handles client (llite) access to multiple OSTs		* '''LOV''' - '''L'''ogical '''O'''bject '''V'''olume - client software layer that handles client (llite) access to multiple OSTs
			* '''LWP''' - '''L'''ight '''W'''eight '''P'''rotocol - a 'lightweight' client co-located on servers that cannot perform replay or recovery
	* '''MDC''' - '''M'''eta'''D'''ata '''C'''lient - client software layer that interfaces to the MDS		* '''MDC''' - '''M'''eta'''D'''ata '''C'''lient - client software layer that interfaces to the MDS
	* '''MDD''' - '''M'''etadata '''D'''evice '''D'''river - MDS software layer that understands POSIX semantics for file access		* '''MDD''' - '''M'''etadata '''D'''evice '''D'''river - MDS software layer that understands POSIX semantics for file access
Line 131:		Line 132:
	* '''OST''' - '''O'''bject '''S'''torage '''T'''arget - storage device that holds the filesystem data (regular data files, not directories, xattrs, or other metadata)		* '''OST''' - '''O'''bject '''S'''torage '''T'''arget - storage device that holds the filesystem data (regular data files, not directories, xattrs, or other metadata)
	* '''PTLRPC''' - '''P'''or'''t'''a'''l''' '''RPC''' - remote procedure call protocol layered over LNet		* '''PTLRPC''' - '''P'''or'''t'''a'''l''' '''RPC''' - remote procedure call protocol layered over LNet
			* '''QMT''' - '''Q'''uota '''M'''aster '''T'''target - Global quota limits are stored in dedicated index files (there is one such index per quota type) on the quota master target (aka QMT)
	* '''SEQ''' - '''SEQ'''uence number of FIDs that are allocated to a client or server for its own exclusive use. The SEQ controller on MDT0000 manages the entire 64-bit SEQ space, and grants large sub-ranges to MDTs and OSTs, which may in turn grant sub-sub-ranges to individual clients.		* '''SEQ''' - '''SEQ'''uence number of FIDs that are allocated to a client or server for its own exclusive use. The SEQ controller on MDT0000 manages the entire 64-bit SEQ space, and grants large sub-ranges to MDTs and OSTs, which may in turn grant sub-sub-ranges to individual clients.

Pjones at 01:00, 25 May 2023

2023-05-25T01:00:24Z

← Older revision		Revision as of 18:00, 24 May 2023
Line 140:		Line 140:
	=== When is the next Major release? ===		=== When is the next Major release? ===

	We plan to tag and build Major releases of Lustre about every ~~six months~~, the [http://lustre.org/roadmap roadmap] gives a high level idea of the current expectations		We plan to tag and build Major releases of Lustre about every year, the [http://lustre.org/roadmap roadmap] gives a high level idea of the current expectations

	=== When is the next Maintenance Release? ===		=== When is the next Maintenance Release? ===

	There are multiple maintenance ~~relesases~~ per year but the exact timing can depend upon several factors (Linux distro releases, OpenZFS releases, important bugfixes). The [http://wiki.opensfs.org/Lustre_Working_Group Lustre Working Group] is the best source of information about the latest status		There are multiple maintenance releases per year but the exact timing can depend upon several factors (Linux distro releases, OpenZFS releases, important bugfixes). The [http://wiki.opensfs.org/Lustre_Working_Group Lustre Working Group] is the best source of information about the latest status

	[[Category: FAQ]]		[[Category: FAQ]]

Adilger: /* What do all these acronyms mean? */ add acronyms from old FAQ page

2020-05-05T01:37:44Z

What do all these acronyms mean?: add acronyms from old FAQ page

← Older revision		Revision as of 18:37, 4 May 2020
Line 99:		Line 99:
	== Acronyms ==		== Acronyms ==
	=== What do all these acronyms mean? ===		=== What do all these acronyms mean? ===
			* '''ACL''' - '''A'''ccess '''C'''ontrol '''L'''ist - more fine-grained file access permission mechanism beyond standard POSIX User/Group/Other permissions
			* '''DLM''' - '''D'''istributed '''L'''ock '''M'''anager - subsystem that manages consistency between multiple client and server access of file/directory data and metadata, also Lustre DLM (LDLM)
	* '''DNE''' - '''D'''istributed '''N'''amespace '''E'''nvironment - feature aggregating multiple MDTs (possibly on multiple MDSes) into a single filesystem namespace		* '''DNE''' - '''D'''istributed '''N'''amespace '''E'''nvironment - feature aggregating multiple MDTs (possibly on multiple MDSes) into a single filesystem namespace
	* '''FID''' - '''F'''ile '''ID'''entifier - unique 128-bit identifier for every object within a single filesystem, made up of a 64-bit sequence (SEQ), a 32-bit object ID (OID), and a 32-bit version field		* '''FID''' - '''F'''ile '''ID'''entifier - unique 128-bit identifier for every object within a single filesystem, made up of a 64-bit sequence (SEQ), a 32-bit object ID (OID), and a 32-bit version field
			* '''FLDB''' - '''F'''ID '''L'''ocation '''D'''ata'''B'''ase - table that maps FID SEQ values/ranges to a specific MDT or OST that holds those objects
			* '''IB''' - '''I'''nfini'''B'''and - high-performance low-latency network interface, alternative to Ethernet
	* '''IDIF''' - OST object '''ID''' '''I'''n '''F'''ID - specific FID range reserved for compatibility with pre-DNE OST objects		* '''IDIF''' - OST object '''ID''' '''I'''n '''F'''ID - specific FID range reserved for compatibility with pre-DNE OST objects
	* '''IGIF''' - '''I'''node and '''G'''eneration '''I'''n '''F'''ID - specific FID range reserved for compatibility from Lustre 1.x MDT inode objects		* '''IGIF''' - '''I'''node and '''G'''eneration '''I'''n '''F'''ID - specific FID range reserved for compatibility from Lustre 1.x MDT inode objects
	* '''IOV''' - '''I'''/'''O''' '''V'''ector		* '''IOV''' - '''I'''/'''O''' '''V'''ector - used to describe file data pages for submission to the network or disk
			* '''LBUG''' - '''L'''ustre '''BUG''' - serious error detected by the kernel code that causes a thread to stop running
			* '''LDLM''' - '''L'''ustre '''D'''istributed '''L'''ock '''M'''anager - see DLM
	* '''LMV''' - '''L'''ogical '''M'''etadata '''V'''olume - client software layer that handles client (llite) access to multiple MDTs		* '''LMV''' - '''L'''ogical '''M'''etadata '''V'''olume - client software layer that handles client (llite) access to multiple MDTs
			* '''LND''' - '''L'''ustre '''N'''etwork '''D'''river - LNet interface to low-level network interfaces such as Ethernet, InfiniBand, etc.
	* '''LNet''' - '''L'''ustre '''Net'''work - Type-agnostic networking layer (TCP/IP on Ethernet, InfiniBand, Omni-Path, Cray Gemini, etc.)		* '''LNet''' - '''L'''ustre '''Net'''work - Type-agnostic networking layer (TCP/IP on Ethernet, InfiniBand, Omni-Path, Cray Gemini, etc.)
	* '''LOD''' - '''L'''ogical '''O'''bject '''D'''evice - MDS software layer that handles access to multiple MDTs and multiple OSTs		* '''LOD''' - '''L'''ogical '''O'''bject '''D'''evice - MDS software layer that handles access to multiple MDTs and multiple OSTs
Line 110:		Line 117:
	* '''MDC''' - '''M'''eta'''D'''ata '''C'''lient - client software layer that interfaces to the MDS		* '''MDC''' - '''M'''eta'''D'''ata '''C'''lient - client software layer that interfaces to the MDS
	* '''MDD''' - '''M'''etadata '''D'''evice '''D'''river - MDS software layer that understands POSIX semantics for file access		* '''MDD''' - '''M'''etadata '''D'''evice '''D'''river - MDS software layer that understands POSIX semantics for file access
	* '''MDS''' - '''M'''eta'''D'''ata '''S'''erver - software service that manages access to filesystem namespace (inodes, paths, permission) requests from the client.		* '''MDS''' - '''M'''eta'''D'''ata '''S'''erver - software service that manages access to filesystem namespace (inodes, paths, permission) requests from the client
	* '''MDT''' - '''M'''eta'''D'''ata '''T'''arget - storage device that holds the filesystem metadata (attributes, inodes, directories, xattrs, etc)		* '''MDT''' - '''M'''eta'''D'''ata '''T'''arget - storage device that holds the filesystem metadata (attributes, inodes, directories, xattrs, etc)
	* '''MGS''' - '''M'''ana'''G'''ement '''S'''erver - service that helps clients and servers with configuration		* '''MGS''' - '''M'''ana'''G'''ement '''S'''erver - service that helps clients and servers with configuration
	* '''MGT''' - '''M'''ana'''G'''ement '''T'''arget - storage device that holds the configuration logs		* '''MGT''' - '''M'''ana'''G'''ement '''T'''arget - storage device that holds the configuration logs
	* '''NID''' - '''N'''etwork '''Id'''entifier - used to uniquely identify a Lustre network endpoint by node ID and network type.		* '''NID''' - '''N'''etwork '''Id'''entifier - used to uniquely identify a Lustre network endpoint by node ID and network type
			* '''OBD''' - '''O'''bject '''B'''ased '''D'''evice - generic term for Lustre devices such as MDT, OST, OSD, LMV, LOV, MDC, OSC
	* '''OFD''' - '''O'''bject '''F'''ilter '''D'''evice - OSS software layer that handles file IO		* '''OFD''' - '''O'''bject '''F'''ilter '''D'''evice - OSS software layer that handles file IO
	* '''OID''' - '''O'''bject '''ID'''entifier - the sub-range of a FID that is allocated by a client within a SEQuence to identify an object		* '''OID''' - '''O'''bject '''ID'''entifier - the sub-range of a FID that is allocated by a client within a SEQuence to identify an object
Line 122:		Line 130:
	* '''OSS''' - '''O'''bject '''S'''torage '''S'''erver - software service that manages access to filesystem data (read, write, truncate, etc)		* '''OSS''' - '''O'''bject '''S'''torage '''S'''erver - software service that manages access to filesystem data (read, write, truncate, etc)
	* '''OST''' - '''O'''bject '''S'''torage '''T'''arget - storage device that holds the filesystem data (regular data files, not directories, xattrs, or other metadata)		* '''OST''' - '''O'''bject '''S'''torage '''T'''arget - storage device that holds the filesystem data (regular data files, not directories, xattrs, or other metadata)
	* '''PTLRPC''' - '''P'''or'''t'''a'''l''' '''RPC''' - A remote procedure call protocol layered over ~~LNET~~		* '''PTLRPC''' - '''P'''or'''t'''a'''l''' '''RPC''' - remote procedure call protocol layered over LNet
	* '''SEQ''' - '''SEQ'''uence number of FIDs that are allocated to a client or server for its own exclusive use. The SEQ controller on MDT0000 manages the entire 64-bit SEQ space, and grants large sub-ranges to MDTs and OSTs, which may in turn grant sub-sub-ranges to individual clients.		* '''SEQ''' - '''SEQ'''uence number of FIDs that are allocated to a client or server for its own exclusive use. The SEQ controller on MDT0000 manages the entire 64-bit SEQ space, and grants large sub-ranges to MDTs and OSTs, which may in turn grant sub-sub-ranges to individual clients.


			For additional glossary terms, see [http://wiki.lustre.org/manual/LustreManual20_HTML/Glossary.html#50438338_pgfId-1006990 ''Glossary''] in the [http://wiki.lustre.org/manual/LustreManual20_HTML/index.html ''Lustre Operations Manual''].

	== Releases ==		== Releases ==

Adilger: /* General */ move updated content from "Frequently Asked Questions (Old Wiki)" page

2020-05-04T21:37:19Z

General: move updated content from "Frequently Asked Questions (Old Wiki)" page

@@ Line 5: / Line 5: @@
 === What is Lustre? ===
 Lustre is a scale-out architecture distributed parallel filesystem.  Metadata services and storage are segregated from data services and storage.
 === What is the difference between a server (MDS/OSS) and a target (MDT/OST)? ===
-A target is the storage device where object (file) or metadata is kept.  The servers service request from clients on their behalf.  Clients do not access storage directly.
+A target is the persistent storage device (HDD or SSD) where object (file) data or metadata is stored.  The servers handle network RPC request on behalf of clients in order to access the files and metadata.  Clients do not access storage directly.
 === Why does lists.lustre.org say "The list overview page has been disabled temporarily"? ===

Adilger: /* What do all these acronyms mean? */ add SEQ and OID, order FID alphabetically

2020-05-04T21:27:03Z

What do all these acronyms mean?: add SEQ and OID, order FID alphabetically

Pjones: /* When is the next Maintenance Release? */

2017-09-05T12:49:59Z

When is the next Maintenance Release?

← Older revision		Revision as of 05:49, 5 September 2017
Line 47:		Line 47:
	=== When is the next Maintenance Release? ===		=== When is the next Maintenance Release? ===

	There are multiple maintenance relesases per year but the exact timing can depend upon several factors (Linux distro releases, OpenZFS releases, important bugfixes). The [http://wiki.opensfs.org/Lustre_Working_Group \| Lustre Working Group] is the best source of information about the latest status		There are multiple maintenance relesases per year but the exact timing can depend upon several factors (Linux distro releases, OpenZFS releases, important bugfixes). The [http://wiki.opensfs.org/Lustre_Working_Group Lustre Working Group] is the best source of information about the latest status

	[[Category: FAQ]]		[[Category: FAQ]]

Pjones: /* When is the next Maintenance Release? */

2017-09-05T12:49:44Z

When is the next Maintenance Release?

← Older revision		Revision as of 05:49, 5 September 2017
Line 47:		Line 47:
	=== When is the next Maintenance Release? ===		=== When is the next Maintenance Release? ===

	There are ~~no longer public Maintenance Releases. Maintenance Releases used to be a public service from~~ the ~~Lustre development team at Intel~~, ~~but they have decided to begin keeping their maintenance branches and~~ releases ~~private~~. ~~Only customers and business partners will have direct access to Intel's maintenance releases~~. ~~For a longer explanation, see [[Retired Release Terminology]]~~.		There are multiple maintenance relesases per year but the exact timing can depend upon several factors (Linux distro releases, OpenZFS releases, important bugfixes). The [http://wiki.opensfs.org/Lustre_Working_Group \| Lustre Working Group] is the best source of information about the latest status

	~~The~~ Lustre ~~Open Source development community remains committed to regularly creating Major Releases along~~ the ~~master branch.~~

	[[Category: FAQ]]		[[Category: FAQ]]

Pjones: /* When is the next Major release? */

2017-09-05T12:39:41Z

When is the next Major release?

← Older revision		Revision as of 05:39, 5 September 2017
Line 43:		Line 43:
	=== When is the next Major release? ===		=== When is the next Major release? ===

	We plan to tag and build Major releases of Lustre every six months, ~~in late March and late September~~.		We plan to tag and build Major releases of Lustre about every six months, the [http://lustre.org/roadmap roadmap] gives a high level idea of the current expectations

	=== When is the next Maintenance Release? ===		=== When is the next Maintenance Release? ===

← Older revision		Revision as of 14:27, 4 May 2020
Line 16:		Line 16:
	=== What do all these acronyms mean? ===		=== What do all these acronyms mean? ===
	* '''DNE''' - '''D'''istributed '''N'''amespace '''E'''nvironment - feature aggregating multiple MDTs (possibly on multiple MDSes) into a single filesystem namespace		* '''DNE''' - '''D'''istributed '''N'''amespace '''E'''nvironment - feature aggregating multiple MDTs (possibly on multiple MDSes) into a single filesystem namespace
			* '''FID''' - '''F'''ile '''ID'''entifier - unique 128-bit identifier for every object within a single filesystem, made up of a 64-bit sequence (SEQ), a 32-bit object ID (OID), and a 32-bit version field
	* '''IDIF''' - OST object '''ID''' '''I'''n '''F'''ID - specific FID range reserved for compatibility with pre-DNE OST objects		* '''IDIF''' - OST object '''ID''' '''I'''n '''F'''ID - specific FID range reserved for compatibility with pre-DNE OST objects
	* '''IGIF''' - '''I'''node and '''G'''eneration '''I'''n '''F'''ID - specific FID range reserved for compatibility from Lustre 1.x MDT inode objects		* '''IGIF''' - '''I'''node and '''G'''eneration '''I'''n '''F'''ID - specific FID range reserved for compatibility from Lustre 1.x MDT inode objects
	* '''FID''' - '''F'''ile '''ID'''entifier - unique 128-bit identifier for every object within a single filesystem, made up of a 64-bit sequence, a 32-bit object ID, and a 32-bit version field
	* '''IOV''' - '''I'''/'''O''' '''V'''ector		* '''IOV''' - '''I'''/'''O''' '''V'''ector
	* '''LMV''' - '''L'''ogical '''M'''etadata '''V'''olume - client software layer that handles client (llite) access to multiple MDTs		* '''LMV''' - '''L'''ogical '''M'''etadata '''V'''olume - client software layer that handles client (llite) access to multiple MDTs
	* '''~~LNET~~''' - '''L'''ustre '''Net'''work - Type agnostic networking layer (TCP/IP on Ethernet, InfiniBand, Omni-Path, Cray Gemini, etc.)		* '''LNet''' - '''L'''ustre '''Net'''work - Type-agnostic networking layer (TCP/IP on Ethernet, InfiniBand, Omni-Path, Cray Gemini, etc.)
	* '''LOD''' - '''L'''ogical '''O'''bject '''D'''evice - MDS software layer that handles access to multiple MDTs and multiple OSTs		* '''LOD''' - '''L'''ogical '''O'''bject '''D'''evice - MDS software layer that handles access to multiple MDTs and multiple OSTs
	* '''LOV''' - '''L'''ogical '''O'''bject '''V'''olume - client software layer that handles client (llite) access to multiple OSTs		* '''LOV''' - '''L'''ogical '''O'''bject '''V'''olume - client software layer that handles client (llite) access to multiple OSTs
Line 32:		Line 32:
	* '''NID''' - '''N'''etwork '''Id'''entifier - used to uniquely identify a Lustre network endpoint by node ID and network type.		* '''NID''' - '''N'''etwork '''Id'''entifier - used to uniquely identify a Lustre network endpoint by node ID and network type.
	* '''OFD''' - '''O'''bject '''F'''ilter '''D'''evice - OSS software layer that handles file IO		* '''OFD''' - '''O'''bject '''F'''ilter '''D'''evice - OSS software layer that handles file IO
			* '''OID''' - '''O'''bject '''ID'''entifier - the sub-range of a FID that is allocated by a client within a SEQuence to identify an object
	* '''OSC''' - '''O'''bject '''S'''torage '''C'''lient - client software layer that interfaces to the OST		* '''OSC''' - '''O'''bject '''S'''torage '''C'''lient - client software layer that interfaces to the OST
	* '''OSD''' - '''O'''bject '''S'''torage '''D'''evice - server software layer that abstracts MDD and OFD access to underlying disk filesystems like ldiskfs and ZFS		* '''OSD''' - '''O'''bject '''S'''torage '''D'''evice - server software layer that abstracts MDD and OFD access to underlying disk filesystems like ldiskfs and ZFS
Line 38:		Line 39:
	* '''OST''' - '''O'''bject '''S'''torage '''T'''arget - storage device that holds the filesystem data (regular data files, not directories, xattrs, or other metadata)		* '''OST''' - '''O'''bject '''S'''torage '''T'''arget - storage device that holds the filesystem data (regular data files, not directories, xattrs, or other metadata)
	* '''PTLRPC''' - '''P'''or'''t'''a'''l''' '''RPC''' - A remote procedure call protocol layered over LNET		* '''PTLRPC''' - '''P'''or'''t'''a'''l''' '''RPC''' - A remote procedure call protocol layered over LNET
			* '''SEQ''' - '''SEQ'''uence number of FIDs that are allocated to a client or server for its own exclusive use. The SEQ controller on MDT0000 manages the entire 64-bit SEQ space, and grants large sub-ranges to MDTs and OSTs, which may in turn grant sub-sub-ranges to individual clients.

	== Releases ==		== Releases ==