Manuale d’uso / di manutenzione del prodotto t2808-90006 del fabbricante HP (Hewlett-Packard)
Vai alla pagina of 104
HP Serviceguard Extended Distance Cluste r for Li nux A.01 .00 Deploy ment Guide Manufacturing P art Number: T2808-9 0006 May 2008 Sec ond Edition.
2 Legal Notices © Copyrig ht 2006-20 08 Hewle tt-Packard Developmen t Company , L.P . Publicat io n Date: 2008 V alid licen se fro m HP require d for po ssessio n, use, or co pying.
3 Cont ent s 1. Disaster T olerance and Recovery in a Ser viceguard Cluster Evaluati ng the Need for Disaster T olerance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 What is a Disaste r T olerant Architecture? . . . . . . . . .
Con tent s 4 Creating a Multiple Di sk Device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2 T o Create and Assemble an MD Device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2 Creating V olume Groups and Configuring VG Exc lusive Activa tion on the MD Mirror .
5 Cont ent s.
Con tent s 6.
7 Printing History The printing date and part number indicate the current edition. The printing date changes when a new edit ion is printed. (Minor corrections and up date s which are in corp orate d at re print d o no t caus e the date t o chan ge.) The pa rt number changes when ex tensive tec hnical chan ges are incorporat ed.
8 HP Prin ting Di vision: Business Critical Computing Business U nit Hewlett-Packard Co. 19111 Pruneridge Ave. Cupertino, CA 95014.
9 Preface This guide i ntroduces t he concept of E xtended Dist ance Clust ers (XDC). It describes how to configure and manage HP Servi ceguard Extended Distance Clusters for Linux and the associated Software RAID functio nality .
10.
11 Preface Related Publicatio ns The followin g docume nts contain addit ional useful informa t ion: • Clusters f or High A vailability: a Primer of HP Solution s , Second Edition .
12.
Disaster T olerance and Recov ery in a S er viceguard Cluster Chapter 1 13 1 Disaster T olerance and Recovery in a Serviceguard Cluster This chap ter introd uces a v ariety of Hew lett-Packard high availab ility cluster tech nologies that provide disaster t olerance fo r your mission-criti cal applicatio ns.
Disaster T olerance and Recov er y in a S er viceguard Cluster Evaluat ing the Need f or Disaster T o lerance Chapter 1 14 Evaluating the Need for Disaster T olerance Disast er tole rance is the abi lity to rest ore applic ations an d data within a reasonable period of time after a disaster .
Disaster T olerance and Recov ery in a S er viceguard Cluster Evaluat ing the Need f or Disaster T o lerance Chapter 1 15 line inoperable as well as the computers. In this case d isaster recovery would be moot, and local failover is probably the more appropriat e level of prot ection.
Disaster T olerance and Recov er y in a S er viceguard Cluster What is a Disaster T olerant Architecture? Chapter 1 16 What is a Disaster T olerant Architecture? In a Serviceguard c luster configuration, high a vailability is ac hieved by using red undant h ardware to el iminat e single points of fail ure.
Disaster T olerance and Recov ery in a S er viceguard Cluster What is a Disaster T olerant Ar chitecture? Chapter 1 17 impact. F or these types o f inst allati ons, and many mo re like t hem, it i s i.
Disaster T olerance and Recov er y in a S er viceguard Cluster Unders tanding T ypes of Disaster T oler ant Clusters Chapter 1 18 Understanding Types of Disas ter T oler ant Clust ers T o protect agai.
Disaster T olerance and Recov ery in a S er viceguard Cluster Understandin g T ypes o f Disaster T olerant Clusters Chapter 1 19 architecture are followed.
Disaster T olerance and Recov er y in a S er viceguard Cluster Unders tanding T ypes of Disaster T oler ant Clusters Chapter 1 20 F igure 1-3 Extended D istance Cluste r In the above configuration the net work and FC links betwee n the data center s a re combined an d se nt over commo n DWDM links .
Disaster T olerance and Recov ery in a S er viceguard Cluster Understandin g T ypes o f Disaster T olerant Clusters Chapter 1 21 F igure 1-4 Two Data Cent er Setup Figure 1-4 sho ws a config urat ion th at is supp orted with separ ate network and FC links between the da ta centers .
Disaster T olerance and Recov er y in a S er viceguard Cluster Unders tanding T ypes of Disaster T oler ant Clusters Chapter 1 22 Also note that the network ing in the configuration sho wn is the minimum. Added network connect ions for additional hea rtbeats are recommended.
Disaster T olerance and Recov ery in a S er viceguard Cluster Understandin g T ypes o f Disaster T olerant Clusters Chapter 1 23 Cluster Extension (CLX) Cluster A Linux CLX cluster is simila r to an HP-UX metropolitan cluster and is a cluster that has alt ernate nodes located in different parts of a city or in nearby cities.
Disaster T olerance and Recov er y in a S er viceguard Cluster Unders tanding T ypes of Disaster T oler ant Clusters Chapter 1 24 Figure 1-5 shows a CLX for a Linux S ervice guard cluster architecture.
Disaster T olerance and Recov ery in a S er viceguard Cluster Understandin g T ypes o f Disaster T olerant Clusters Chapter 1 25 Benefi ts of CLX • CLX offe rs a more resili ent so lution t han Ext .
Disaster T olerance and Recov er y in a S er viceguard Cluster Unders tanding T ypes of Disaster T oler ant Clusters Chapter 1 26 • D isk resy nchron ization is indep endent o f CPU failure (th at is, if the hosts a t the primar y site fail b ut the disk rema ins up, the disk kno ws it does not h ave to be resync hronized).
Disaster T olerance and Recov ery in a S er viceguard Cluster Understandin g T ypes o f Disaster T olerant Clusters Chapter 1 27 "objective" can be set for the recove ry point such that if data is upda ted for a pe rio d le ss th an the ob je cti ve, auto mat ed f ai love r can occur and a package w ill start.
Disaster T olerance and Recov er y in a S er viceguard Cluster Unders tanding T ypes of Disaster T oler ant Clusters Chapter 1 28 F igure 1-6 Continental Cluster Contine ntalclus ters p rovi des the flexi bility to wo rk with any data replica tion mec hanism.
Disaster T olerance and Recov ery in a S er viceguard Cluster Understandin g T ypes o f Disaster T olerant Clusters Chapter 1 29 Benefi ts of Continenta lclusters • Y ou can virtua lly build data center s anywhere and still have the data centers prov ide disast er tolerance for ea ch other .
Disaster T olerance and Recov er y in a S er viceguard Cluster Unders tanding T ypes of Disaster T oler ant Clusters Chapter 1 30 replica te the data between t wo data cente rs . HP provid es a suppo rted inte gration toolki t for Ora cle 8i St andby DB in t he Enterprise Cluster Management T oolkit (ECMT).
Disaster T olerance and Recov ery in a S er viceguard Cluster Understandin g T ypes o f Disaster T olerant Clusters Chapter 1 31 T able 1-1 Comparison of Disas ter T olerant Cluster Solutions Attributes Extended Distance Cluster CLX Contin entalclusters (HP-UX only) Key Be nefit Excelle nt in “no rmal” operations , and part ial failure .
Disaster T olerance and Recov er y in a S er viceguard Cluster Unders tanding T ypes of Disaster T oler ant Clusters Chapter 1 32 Key Limita tion N o abilit y to che ck the state o f the da ta before starting up th e appli cat ion. If the volume group (v g) can be act iva ted, th e appli cat ion will b e started.
Disaster T olerance and Recov ery in a S er viceguard Cluster Understandin g T ypes o f Disaster T olerant Clusters Chapter 1 33 Maximum Distance 100 K ilome ters Shorte st of the di stan ces betwe en: •C l u s t e r n e t w o r k laten cy (not to exceed 200 ms).
Disaster T olerance and Recov er y in a S er viceguard Cluster Unders tanding T ypes of Disaster T oler ant Clusters Chapter 1 34 Applicatio n F ailover type Automatic (no manual interv ention requi r ed). Automatic (no manual intervention required). Semi-a utoma tic ( user must “pus h the button” to init iate recovery).
Disaster T olerance and Recov ery in a S er viceguard Cluster Understandin g T ypes o f Disaster T olerant Clusters Chapter 1 35 Data Replicatio n Link Dark Fiber Dark Fiber Continuous Access over IP .
Disaster T olerance and Recov er y in a S er viceguard Cluster Unders tanding T ypes of Disaster T oler ant Clusters Chapter 1 36 DTS Software/ Licen ses Required SGLX + XDC SGLX + CLX XP or CLX EV A .
Disaster T olerance and Recov ery in a S er viceguard Cluster Disaster T olerant Ar chitecture Guidelines Chapter 1 37 Disaster T olerant Architec ture Guidelines Disa ster t olera nt arc hitec tures repr esen t a shif t away from the mass ive central data centers and tow ards more distributed data processing facilitie s .
Disaster T olerance and Recov er y in a S er viceguard Cluster Disaster T olerant Ar chitecture Guid elines Chapter 1 38 Protecting Data through Replication The mo s t sig nifi cant l osses d uring a d isas ter are the lo ss of ac cess to data, an d the lo ss of d ata it self.
Disaster T olerance and Recov ery in a S er viceguard Cluster Disaster T olerant Ar chitecture Guidelines Chapter 1 39 depending on the volume of data. So me applicat ions, depending on the role they play in the bus iness, ma y need to have a faster recovery t ime, within hours or eve n minutes.
Disaster T olerance and Recov er y in a S er viceguard Cluster Disaster T olerant Ar chitecture Guid elines Chapter 1 40 F igure 1-7 Physical Data Replication MD Software RAID is an example of phys ic.
Disaster T olerance and Recov ery in a S er viceguard Cluster Disaster T olerant Ar chitecture Guidelines Chapter 1 41 • Th e logical order of data w rites is not al ways maintained in synchronou s repl ication.
Disaster T olerance and Recov er y in a S er viceguard Cluster Disaster T olerant Ar chitecture Guid elines Chapter 1 42 • Because there a re multiple read devices , tha t is , the node ha s access to both copies of data , there ma y be imp r ovements in read performance .
Disaster T olerance and Recov ery in a S er viceguard Cluster Disaster T olerant Ar chitecture Guidelines Chapter 1 43 F igure 1-8 Logical Data R eplication Advantages of using logical replic ation are : • The distance betwee n nodes is limited only by the networking technol ogy .
Disaster T olerance and Recov er y in a S er viceguard Cluster Disaster T olerant Ar chitecture Guid elines Chapter 1 44 • I f the pri mary data base fails and is co rrupt, which res ults in t he repli ca t aking over , t hen the p ro cess for r est ori ng t he pr im ary dat abas e so t hat i t can be use d as th e re plic a is co mple x.
Disaster T olerance and Recov ery in a S er viceguard Cluster Disaster T olerant Ar chitecture Guidelines Chapter 1 45 F igure 1-9 Alternative P ower Sources Housing re mote n odes in anothe r buildin.
Disaster T olerance and Recov er y in a S er viceguard Cluster Disaster T olerant Ar chitecture Guid elines Chapter 1 46 Disas ter T olerant Loc al Area N etwo rking Ethernet networks can also be used.
Disaster T olerance and Recov ery in a S er viceguard Cluster Disaster T olerant Ar chitecture Guidelines Chapter 1 47 Disaster T olerant Cluster Limitations Disast er tolera nt clusters h ave limitatio ns, some of which ca n be mitigated b y good planni ng.
Disaster T olerance and Recov er y in a S er viceguard Cluster Managi ng a Disaster T olerant En vir onment Chapter 1 48 Managing a Di saster T olera nt Environment In addition to the changes in hardware and so ft ware to create a disaster tolerant archite cture, t here are also c hanges in the w ay you mana ge the environmen t.
Disaster T olerance and Recov ery in a S er viceguard Cluster Managi ng a Disaster T olerant En vir onment Chapter 1 49 Even if recovery is automated, you may choose to , or need to recover from some types of di sasters with manual r ecovery .
Disaster T olerance and Recov er y in a S er viceguard Cluster Additional Disaster T olerant Solutions Inf ormation Chapter 1 50 Additional Disaster T olerant Solutions Information On- line v ersio ns o f HA doc ument atio n are a vail able a t http:// docs.
Building an Extended Distance Cluster Using Ser viceguard and Software RAID Chapter 2 51 2 Building an Extended Distance Cluster Using Serviceguard and Software RAID Simple Se rviceguard clusters are .
Building an Extended Distance Cluster Using Ser vic eguard and Software RAID T ypes of Data Link f or Stor age and Netwo rking Chapter 2 52 Types of Data Link for Storage and Networking Fib re Channel.
Building an Extended Distance Cluster Using Ser viceguard and Software RAID T wo Data Center and Qu orum Service Location Ar chitectures Chapter 2 53 Two Data Center and Quorum Service Location Archit.
Building an Extended Distance Cluster Using Ser vic eguard and Software RAID T wo Data Center and Quorum Service Locat ion Ar chitectures Chapter 2 54 • Fibre Channel Direct F a bric Attach (DF A) i.
Building an Extended Distance Cluster Using Ser viceguard and Software RAID T wo Data Center and Qu orum Service Location Ar chitectures Chapter 2 55 F igure 2-1 Two Data Centers and Third Location wi.
Building an Extended Distance Cluster Using Ser vic eguard and Software RAID T wo Data Center and Quorum Service Locat ion Ar chitectures Chapter 2 56 There are no requirement s for the dis t ance bet.
Building an Extended Distance Cluster Using Ser viceguard and Software RAID Rules f or Separate N etwork and D ata Links Chapter 2 57 Rules for Separate Network and Data Lin ks • Th ere must be less than 200 milli seconds o f latenc y in t he netwo rk betwe en the d ata ce nter s.
Building an Extended Distance Cluster Using Ser vic eguard and Software RAID Guidelines on D W DM Links fo r Network and Data Chapter 2 58 Guidelines on DWDM Links for Netwo rk and Data • Th ere must be less than 200 milli seconds o f latenc y in t he netwo rk betwe en the d ata ce nter s.
Building an Extended Distance Cluster Using Ser viceguard and Software RAID Guidelines on D WDM Links for Netw ork and Data Chapter 2 59 • Fibre Channel switc hes must be used in a DWDM configura tion; Fibre Cha nnel hubs are not su pport ed. Direct F abric Attach mode must be used for the ports connected to the DWDM link.
Building an Extended Distance Cluster Using Ser vic eguard and Software RAID Guidelines on D W DM Links fo r Network and Data Chapter 2 60.
Configur ing your En vironment f or Software RAID Chapter 3 61 3 Configuring your Environment for Software RAID The previ ous c hapters di scussed conce ptual informatio n on disaster tolerant a rchitect ures and proced ural in formation on creatin g an extended dista nce cl uster .
Configuring your Environment f or Software RAID Under standing Softwar e RAID Chapter 3 62 Underst an d ing So ftwa re R AI D Redundant Array of Independent Disks (RAID) is a mechanism that provides storage fault tolerance and , occasiona lly , better performance .
Configur ing your En vironment f or Software RAID Installing t he Extended Distance Cluster Software Chapter 3 63 Installing t he Extended Distance Cluster Software This se ction disc usses the s upported opera ting sys tems , prereq uisite s and the pro cedures for instal ling the Exten ded Distance C luster softwar e.
Configuring your Environment f or Software RAID Installing the Ext ended Distance Cluster Sof tware Chapter 3 64 Compl ete the fo llow ing p roce dur e to in stall XDC: 1. Insert the pr oduct CD into th e driv e and mou n t the CD . 2. Open the command l ine inte rface.
Configur ing your En vironment f or Software RAID Installing t he Extended Distance Cluster Software Chapter 3 65 In the outpu t, the produ ct name , xdc -A.01 .00-0 will be listed. The presence of th is file verifies that the instal latio n is succe ssful .
Configuring your Environment f or Software RAID Configur ing the En vironment Chapter 3 66 Configuring the Environment After setting up the hardw are as described in the Extend ed Distance Cluster Arc.
Configur ing your En vironment f or Software RAID Configur ing the En vironment Chapter 3 67 that a re of ide n tical si zes . Diff erences i n disk se t size re sults in a mirror being cre ated of a s ize equal to the smal ler of the t wo disks . Be sure to create the mirror us ing the p ersistent d evice name s of th e component devices .
Configuring your Environment f or Software RAID Configur ing the En vironment Chapter 3 68 • Ensure that th e Quorum Serv er link i s close to th e Ethernet l inks in your setup. In cases of fail ures of all Et hernet and Fibre channel links, the node s can easily access the Quorum Server for arbitration.
Configur ing your En vironment f or Software RAID Configuring Mult iple Pa ths to Storage Chapter 3 69 Configuring Multiple P aths to Sto rag e HP require s that yo u configu re multi ple pa ths to the stora ge de vice using the QLogic HBA driver as it has inbuilt multipath capabilitie s.
Configuring your Environment f or Software RAID Configur ing Multiple P aths t o Stora ge Chapter 3 70 The QLogic cards are configured t o ho ld up any disk access and essential ly hang for a time period whic h is greater than the c luster reformation time w hen access to a disk is l ost.
Configur ing your En vironment f or Software RAID Using Persi stent Device Names Chapter 3 71 Using P ersistent Device Names When there i s a disk rel ated failure a nd subseq uent reboot, th ere is a possi bility tha t the d evi ces are rena med. Linux name s di sks i n th e orde r they are f ound.
Configuring your Environment f or Software RAID Creating a Multiple Disk De vice Chapter 3 72 Creating a Multiple Disk Device As mention ed earlier , the first step for enab ling Softw are RAID in your environment is to create the Multiple Disk (MD) device using two underlying componen t disks.
Configur ing your En vironment f or Software RAID Creating a Multip le Disk Device Chapter 3 73 2. Assemble the MD device on the o ther node by running the following command: # mdadm -A -R /dev/md 0 /dev/h pdev/sde1 /dev/hp dev/sd f1 3.
Configuring your Environment f or Software RAID Creating V olume Gr oups and Configuring V G Exc lusive Activ ation on the MD Mirr or Chapter 3 74 Creating V olume Groups and Conf iguring VG Exclusive Ac tivation on the MD Mi rror Once you cre ate the MD mirror d evice , you ne ed to create vo lume g roups and logical v olumes o n it.
Configur ing your En vironment f or Software RAID Creating V olume Gr oups and Con figuring V G Exclusive Ac tivati on on the MD Mirr or Chapter 3 75 Found d uplica te PV 9w 3TIxKZ6l FRqWUmQm9 tlV5nsdU kTi4i: using /dev/sd e not /dev/sdf Wi th this error , you cannot create a new volume group on /dev/md0 .
Configuring your Environment f or Software RAID Configur ing the P ac kage Contr ol Script and R AID Configur ation File Chapter 3 76 Configuring the P ackage Control Script an d RAID Configuration F .
Configur ing your En vironment f or Software RAID Configur ing the P ac kage Contr ol Script and RAI D Configuration File Chapter 3 77 # Specify the method of activation and deactivation for md.
Configuring your Environment f or Software RAID Configur ing the P ac kage Contr ol Script and R AID Configur ation File Chapter 3 78 T o Edit the XDC_CONFIG FILE parameter In additio n to modifying the DATA _REP variable, you must also s et XDC_CON FIG_FI LE to spec ify th e raid.
Configur ing your En vironment f or Software RAID Configur ing the P ac kage Contr ol Script and RAI D Configuration File Chapter 3 79 more time elapse s than w hat is sp ecifie d for RPO_TARG ET , the package is prevente d from start ing on the remote no de (assuming that the node stil l has acce ss only to its o wn half of the mirror).
Configuring your Environment f or Software RAID Configur ing the P ac kage Contr ol Script and R AID Configur ation File Chapter 3 80 F or example, let us as sume tha t the dat a storag e links in Figu re 1-4 fail befo re the hear tbeat lin ks fail .
Configur ing your En vironment f or Software RAID Configur ing the P ac kage Contr ol Script and RAI D Configuration File Chapter 3 81 Now conside r an XDC configurat ion such as that shown in F igure 1-3 (DWDM links betw een data cent ers).
Configuring your Environment f or Software RAID Configur ing the P ac kage Contr ol Script and R AID Configur ation File Chapter 3 82 Again, if the network is set up in suc h a way that when the links.
Configur ing your En vironment f or Software RAID Configur ing the P ac kage Contr ol Script and RAI D Configuration File Chapter 3 83 • RAID_MONITOR_INTERV AL This parameter defines the time interv.
Configuring your Environment f or Software RAID Configur ing the P ac kage Contr ol Script and R AID Configur ation File Chapter 3 84.
Disaster Scenarios and Their Handling Chapter 4 85 4 Disaste r Sc enarios and Their Handling The pr evious cha pters provi ded informat ion on depl oying Soft ware RAID in your en vironmen t. In th is chapter , you w ill find i nformation on ho w Softw are RAID addr esses variou s disaster sc enarios .
Disaster Scenar ios and Their Handling Chapter 4 86 The fo llow ing ta ble li sts al l the disas ter sce nario s that are h andle d by the Ext ended Dist ance Clust er softw are . All the scenari os assume that the s etup is th e same a s the one describ ed in “Exte nded Dis tance Clust ers” o n page 18 of this d ocument .
Disaster Scenarios and Their Handling Chapter 4 87 A packag e (P1) is runni ng on a node ( Node 1 ). The package uses a mirror (md0) that consis ts of two stor ag e com pone nt s - S1 (local to Node 1 - /dev/hpdev/mylink-sde ) and S2 (local to Node 2).
Disaster Scenar ios and Their Handling Chapter 4 88 A packag e (P1) is runni ng on a node ( Node 1 ). The package uses a mirror (md0) that consis ts of two stor ag e com pone nt s - S1 (local to Node 1 - /dev/hpdev/mylink-sde ) and S2 (local to Node 2) Data center 1 that c onsists of Node 1 and P1 ex periences a failu re.
Disaster Scenarios and Their Handling Chapter 4 89 This i s a mult iple fail ure scena rio wh ere the f ailur es occur i n a particular se quence in the con figuratio n that correspo nds to figure 2 where Ethernet and FC l inks do no t go over DWDM . The pac kage (P1) is running o n a node (N1) .
Disaster Scenar ios and Their Handling Chapter 4 90 This i s a mult iple fail ure scena rio wh ere the f ailur es occur i n a particular se quence in the con figuratio n that correspo nds to figure 2 where Ethernet and FC l inks do no t go over DWDM .
Disaster Scenarios and Their Handling Chapter 4 91 This fa ilure is t he same as the previo us f ailu re excep t tha t the packag e (P 1) i s c onfi gur e d wit h RPO_TARGET set to 60 seconds . In thi s case, initia lly the packag e (P1) is runni ng on N 1.
Disaster Scenar ios and Their Handling Chapter 4 92 In this case, the package (P1) runs with RPO _TARGET set to 60 seco nds. P acka ge P1 is runnin g on node N1. P1 uses a mirro r md0 consisting of S1 (loc al to node N1, for exa mple /dev/hpdev/mylink-sde ) and S2 (local to no de N2).
Disaster Scenarios and Their Handling Chapter 4 93 This s cenario is a n extensi on of the pre vio u s fail ure s cena rio. In the previo us sce nario, when the package f ails ov er to N2, it does not start as the value of RPO_TARGET would have been exceed ed.
Disaster Scenar ios and Their Handling Chapter 4 94 In this case, the package (P1) runs with RPO -TARGET set to 60 seco nds. In thi s case, initia lly the pack age ( P1) i s r unn ing on n ode N1. P1 uses a mirro r md0 consisting of S1 (loc al to node N1, for exa mple /dev/hpdev/mylink-sde ) and S2 (local to no de N2).
Disaster Scenarios and Their Handling Chapter 4 95 In thi s case, initia lly the pack age ( P1) i s r unn ing on n ode N1. P1 uses a mirro r md0 consisting of S1 (loc al to node N1, for exa mple /dev/hpdev/mylink-sde ) and S2 (local to no de N2). The first failure occu rs with all Eth ernet links between the two data center s failing .
Disaster Scenar ios and Their Handling Chapter 4 96.
Managing an MD De vice Appendix A 97 A Managing an MD Devi ce This chapte r includes addition al inf ormation on how to manage t he MD device . F or the latest information on how to manage and MD device, see The S oftware-RAID HOWTO ma nual availa ble at: http:// www.
Managing an MD De vice Vie wing the Status of the MD Device Appendix A 98 V iewing the Status of the MD Device After creating an MD device, yo u can view its status. By doing so , you can remain info rmed of wheth er the device is clean, up and running , or i f there ar e an y erro r s.
Managing an MD De vice Stoppin g the MD Device Appendix A 99 Stopping the MD Device After you creat e an MD dev ice, it begins to run. Y ou need to s top the device an d add the config uration into the rai d.
Managing an MD De vice Starting the MD Device Appendix A 100 Starting the MD Device After you c reate an MD devi ce, yo u would need t o s top and start t he MD device to en sure that it is active. Y ou woul d not need to start the MD device in a ny othe r scenari o as this is han dled by the XDC s oftware.
Managing an MD De vice Removing and Adding an MD Mirr or Co mponent Disk Appendix A 101 Removi ng and Adding an MD Mirror Component Di sk There are ce rtain failure scenarios , where y ou would need t o manually remove the mirror component of an MD d evice and add it again later .
Managing an MD De vice Remo ving and Add ing an MD M irror C omponent D isk Appendix A 102 Example A-3 Removing a fail ed MD component d isk from /dev /md0 array T o remove a f ailed MD co mpon ent d .
Index 103 A asynchronous data replication , 39 C clus te r exten ded di stance , 22 FibreCha nnel , 52 metropolitan , 23 wide area , 27 cluster main tenan ce , 49 configuring , 46 disaster t olerant E.
Index 104 persistent device names , 66 physica l data replicati on , 39 power sources redundant , 44 Q QLogic cards , 70 R RAID Mo nitor ing Se rvice Configure , 78 raid.
Un punto importante, dopo l’acquisto del dispositivo (o anche prima di acquisto) è quello di leggere il manuale. Dobbiamo farlo per diversi motivi semplici:
Se non hai ancora comprato il HP (Hewlett-Packard) t2808-90006 è un buon momento per familiarizzare con i dati di base del prodotto. Prime consultare le pagine iniziali del manuale d’uso, che si trova al di sopra. Dovresti trovare lì i dati tecnici più importanti del HP (Hewlett-Packard) t2808-90006 - in questo modo è possibile verificare se l’apparecchio soddisfa le tue esigenze. Esplorando le pagine segenti del manuali d’uso HP (Hewlett-Packard) t2808-90006 imparerai tutte le caratteristiche del prodotto e le informazioni sul suo funzionamento. Le informazioni sul HP (Hewlett-Packard) t2808-90006 ti aiuteranno sicuramente a prendere una decisione relativa all’acquisto.
In una situazione in cui hai già il HP (Hewlett-Packard) t2808-90006, ma non hai ancora letto il manuale d’uso, dovresti farlo per le ragioni sopra descritte. Saprai quindi se hai correttamente usato le funzioni disponibili, e se hai commesso errori che possono ridurre la durata di vita del HP (Hewlett-Packard) t2808-90006.
Tuttavia, uno dei ruoli più importanti per l’utente svolti dal manuale d’uso è quello di aiutare a risolvere i problemi con il HP (Hewlett-Packard) t2808-90006. Quasi sempre, ci troverai Troubleshooting, cioè i guasti più frequenti e malfunzionamenti del dispositivo HP (Hewlett-Packard) t2808-90006 insieme con le istruzioni su come risolverli. Anche se non si riesci a risolvere il problema, il manuale d’uso ti mostrerà il percorso di ulteriori procedimenti – il contatto con il centro servizio clienti o il servizio più vicino.
