+ All Categories
Home > Documents > Standards in DPM

Standards in DPM

Date post: 10-Feb-2016
Category:
Upload: gabi
View: 33 times
Download: 0 times
Share this document with a friend
Description:
Standards in DPM. Outline. Disk Pool Manager (DPM) Grid Data Management Why Standards POSIX Access (NFS4.1) HTTP / WebDAV. Disk Pool Manager (DPM). DPM Main Goals. Provide lightweight “grid enabled” storage Manage space on distributed disk servers Manage a hierarchical namespace - PowerPoint PPT Presentation
Popular Tags:
37
Standards in DPM
Transcript
Page 1: Standards in DPM

Standards in DPM

Page 2: Standards in DPM

Outline

• Disk Pool Manager (DPM)• Grid Data Management• Why Standards• POSIX Access (NFS4.1)• HTTP / WebDAV

Page 3: Standards in DPM

Disk Pool Manager (DPM)

Page 4: Standards in DPM

DPM Main Goals

• Provide lightweight “grid enabled” storage• Manage space on distributed disk servers• Manage a hierarchical namespace• Expose interfaces for– Space Management

(socket,SRM1.1,SRM2.1,SRM2.2)– Remote data access (gridFTP,HTTP/HTTPS)– POSIX like access (rfio)

Page 5: Standards in DPM

DISK NODEDISK NODE

DPM Architecture

HEAD NODE

DISK NODES

CLIENT

METADATADPMDPNS SRM

RFIO HTTP XROOT

GRIDFTPCLI PYTHON C API NFS 4.1

DIRECT DATA

ACCESS

NFS 4.1

Page 6: Standards in DPM

DPM Architecture

HEAD NODE

DPMDPNS SRM

METADATAFile and directory metadata operations

File access requests, interaction with Disk Node

daemons

Storage management via SRM clients

NFS 4.1

Page 7: Standards in DPM

DISK NODEDISK NODE

DPM Architecture

DISK NODES

RFIO HTTP XROOT

GRIDFTP NFS 4.1

POSIX like access

GET / PUT operations (Apache Web Server)

File upload / download using GSI over FTP

POSIX like access

Page 8: Standards in DPM

DPM Main Features• Separation of data and metadata• Hierarchical namespace

– /dpm/<domain>/home/<vo>• Strong security

– X509 & KRB5 (auth), VOMS & Virtual IDs (authz)• UNIX like commands

– dpns-ls, dpns-mkdir, dpns-<usual-command-here>• Database centric

– Support for MySQL, Oracle and PostgreSQL, easily load balanced• SLC4/5 and Solaris for server side, Debian clients,

prototype for MacOSX clients

Page 9: Standards in DPM

DPM Further Details

• Written mostly in C• Statistics from GStat– https://gstat-wlcg.cern.ch/gstat/stats/– Over 200 grid sites use DPM

• Largest deployment: 1.2PB• More Information– https://svnweb.cern.ch/trac/lcgdm/wiki/Dpm

Page 10: Standards in DPM

Grid Data Management

Page 11: Standards in DPM

Grid Data Management

• Heterogeneous environment• > 250 sites in total• DPM is one of multiple storage elements– Others are CASTOR, dCache, STORM/GPFS, …

• SRM is the only “standard” protocol– Helps, but only for space management– And even here implementations differ

• No standard protocol for data access– rfio, dcap(++), xroot

Page 12: Standards in DPM

Grid Data Management

CLIENT INFOSYS

DCACHE

DCAP / DCAP++

GRIDFTP HTTP / WEBDAV

XROOT

CASTOR

RFIO

GRIDFTP

XROOT

DPM

RFIO

GRIDFTP HTTP

XROOT

...

GFAL

• Clients need knowledge of storage backend type

• Complex, hard to deploy and maintain

BDIIToA…

Page 13: Standards in DPM

Grid Data Management

• Library dependency issues• Requirement of user interfaces (UIs)– Entry points to the grid– Maintained by experts

• Very hard to use “standard” distributions– Even transition from SLC4 to 5 is problematic

• Validation takes a long time

Page 14: Standards in DPM

Why Standards?

• Accessibility– Not limiting access to OS X version Y with library Z

• Validation– Using common validation and test tools

• Stability– Evolution discussed in a wide group

• Ease of implementation– Sharing of experiences, common code base

• No vendor lock-in• …

Page 15: Standards in DPM

POSIX Data Access & NFS 4.1

Page 16: Standards in DPM

POSIX Data Access• We have some specific requirements• Strong authentication

– Ideally using X509 certificates• Support for clustered filesystems

– Separation of data and metadata access• (Global?) Hierarchical namespace• Performance (even in WAN)

NFS4.1 offers all of this… and more• We now have a standard we can use• It’s just POSIX, no need for any additional library

Page 17: Standards in DPM

Some NFS History• NFS2 in 1989 (RFC 1094)

– NFS1 was Sun internal• NFS3 in 1995 (RFC 1813)

– Large file support (64bit)– Performance enhancements (transfer buffer, num round trips)

• NFS4 in 2003 (RFC 3010)– Better WAN performance– Strong security– Locking– Delegations– Callbacks– Backwards compatible extensions

• NFS4.1 in 2010 (RFC 5661)– Main feature is parallel NFS

Page 18: Standards in DPM

NFS 4.1

• IETF Standard (RFC 5661)• Different in nature from previous versions• Parallelism is the key word– No single server bottleneck– Meets needs of HPC and clustered systems

• Supported by major vendors• Let’s look a bit in detail…

Page 19: Standards in DPM

NFS 4.1 Overview

CLIENT

METADATA SERVER

DATASERVER

CONTROL PROTOCOL(undefined)

pNFS Protocol Storage Access Protocol(Layouts)

DATASERVER

Callbacks

Page 20: Standards in DPM

NFS 4.1 Feature 1 – Unified Protocol

• One protocol, one port (2049)• Previous versions required additional

protocols– mount, lock, status, …

Page 21: Standards in DPM

NFS 4.1 Feature 2 – Strong Auth(z)• Based on RPCSEC_GSS (RFC 2203)• Support for multiple security mechanisms

– KRB5 is mandatory, negotiation is in the protocol– Working to have X509 support (probably via globus GSSAPI plugin)

• String based identities• Basic permissions + ACLs• Example: Linux Client

KERNEL SPACE

USER SPACE

KERNEL CLIENT

RPC.GSSDCREDENTIAL STORE

MECHGLUE

1

22

3

sec contextnegotiation

system call

Page 22: Standards in DPM

NFS 4.1 Feature 3 – Bulk Operations

• Protocol defines only two procedures– NULL and COMPOUND

• COMPOUND procedure holds Operations– Open, Read, Write, Close, …

• Much less round trips– Better performance, especially over WAN

Page 23: Standards in DPM

NFS 4.1 Feature 4 - Sessions

• Decouples transport (connection) from client• Persistent state on the server– Locks, opens, delegations, layouts, …

• Multiple connections per session (from the same client)

• Multiple sessions per client• Exactly Once Semantics (EOS)– Even in failure/recovery, thanks to reply cache

Page 24: Standards in DPM

NFS 4.1 Feature 5 - Delegations

• Given for files and directories• Moves part of the logic from server to client– Regarding access permissions, …

• Multiple types of delegations• Can be recalled (via callback)– On conflicting request from other client

• Improved Performance– Less round trips

Page 25: Standards in DPM

NFS 4.1 Feature 6 - Layouts

• Describe how to access data in storage– Multiple storage protocols supported• File, Object (RFC 5663), Block (RFC 5664)• Striping available with pNFS File Layout

• Clients must request layout to MD server• Storage servers refuse access if request does

not match layout• Layouts can be recalled, via callback– Ex: change in access permissions

Page 26: Standards in DPM

NFS 4.1 Feature 7 – Multi Server Namespace

• Namespace spawning multiple domains• Servers redirect clients when data is not local– Redirection is the key word here

• Can also be used to provide clients with alternative locations

CLIENT

NFS SITE A

NFS SITE B

1

2

open (/grid/siteB/myFile)

open (/grid/siteB/myFile)

NFS4ERR_MOVED ( fs_locations )

Page 27: Standards in DPM

NFS 4.1 Additional Goodies

• Clients provided by industry– Linux, Solaris, Windows

• Free client caching– It’s just there… we benefit from experts

implementing caching in the OS• Support from major industry vendors– Netapp, Panasas, IBM, Oracle, EMC– Waiting for wide client availability– dCache also has support for NFS4.1

Page 28: Standards in DPM

NFS 4.1 Client Availability

• Linux since 2.6.32– pNFS part coming with 2.6.36– pNFS builds in Fedora 12, 13• We also keep a debian one

– pNFS expected in RHEL 6.1• Solaris driver available (but not shipped yet)• Windows driver available

Page 29: Standards in DPM

DPM NFS4.1 Implementation

• Aiming for a basic prototype by end of August• Additional frontend in the DPM Head Node• Possible to reuse existing implementations– Benefit from implementing a standard– Kernel Space: Linux spnfs– User Space: Ganesha (CEA)

Page 30: Standards in DPM

DPM NFS4.1 Implementation

• Based on sPNFS– A prototype pNFS server implementation for Linux– Done as a proof of concept, but good for us to use as a

starting point• Still, a lot of the code can be reused– We integrate the name server requests with the DPM API

NFSDMODULE

DPM NFS DAEMON

KERNEL SPACE

USER SPACE

PIPEFS

DPM API

Page 31: Standards in DPM

RPC.GSSDUSER APP

NFS KERNEL CLIENT

NFS KERNEL SERVER

RPC.SVCGSSD DPM NFS DAEMON

DPM APIDPM DB

CLIENT SIDE SERVER SIDE

KERNEL SPACE

USER SPACE

DPM NFS 4.1 Picture

Page 33: Standards in DPM

Data Transfer

Page 34: Standards in DPM

HTTP (s)

• DPM already supports HTTP (s)– As a transfer protocol

• Easy authentication / authorization– Newer versions of openssl with X509 proxy

support make this even easier• Implemented as an apache module or cgi• Firewall friendly• Clients? They are everywhere…

Page 35: Standards in DPM

WebDAV

• Extensions to HTTP 1.1 for document management (RFC 2518)

• Enables wide collaboration– Locking– Namespace management (copy, move, …)– Metadata / properties on files

• Maybe not so interesting for HEP users– But very popular within other communities– dCache has had very good feedback on it

• Implementation not yet scheduled, but in the plan

Page 36: Standards in DPM

Conclusions

• Our environment is not standards friendly• Standard protocols exist today fitting all our use

cases• Benefits for users, developers, admins– Usability, maintainability, evolution

• DPM will continue focusing on standards– And will soon use them for all our use cases

• Ongoing work also within the EMI data management group in the same direction

Page 37: Standards in DPM

?


Recommended