Chapter 17: Recovery System

صفحه 1:
Cheaper 07: Qevpvery Opsew 

صفحه 2:
+ ‏وان‎ 0۳: Racer Oyster ‎OkssFiccion‏ ساد ‎Gtorage ‏سس‎ ‎Recovery od ‏رو‎ ‎LoxpBased Recovery ‎Gkodow Pata ‎Recovery Dik Oowured Treceuniow OPPer Oonnewect ‎(Poture with Loss oF Ovavohathe Store Odcred Recovery ‏تج هل‎ @ORIEG Revovery Okprikee ‎Rewoe Backup Oysters ‎@eedrwr Gyetre Oocowytr, Oe. ne ©Sbervehnts, Cork ced Cnakershe 

صفحه 3:
ح ات0 سول + ۲ ‏موب‎ Poke? eee merce cee ee eee ee ee ee te ‏مج‎ مس ره مه ها اس مرو لو ‎Cyetew errors! he‏ © ‎dexfock)‏ ری مج و مت صا جل همه سوت اوه و سور او و سا سوه تلو مسق ۲ او و مرو با ی اس مب بو وی و ارو زا اه و بو تمه هط تاه روا موه ما مرو له( یب ‎Baker destroys df or pat of debe‏ باس مور اه میاه تسوا 0 ۳ ‎pion‏ © Desiring is weaved tp be detevicble! disk drives use chevkouvs ty detent ‏یله‎ |۱۳ Gyetre Oocowytr, Oe. ne ©Sbervehnts, Cork ced Cnakershe 

صفحه 4:
+ Revovery ‏و‎ له ری لعج وا هه ول ترس ۲ اه( ال رالد له توت شمسا ©) Come oP this chapter © Revvery cherthws hove tw ports 6 Qrtces toed durtog aero trocsaniva processtoy io Poure Porch iPorwoticg exists to recover Pro Phares واه و وا عون ‎the dotubase‏ رو وا ولو و له وله ‎(Brticos‏ ‏رات له ری ,وود سور ‎Od. ne ©Sbervehnts, Cork ced Cnakershe‏ ,تن 6 تیه 

صفحه 5:
+ Ouray Onoter © Ockile strroce! © does wt sue syste rushes © ‏من رم‎ eur, cache ‏رو‎ ‎© Ocak serene! © sunives aysiew rushes © exnopkes! cok, tore, Phok ww, cexruokile (botery barked uw) REO © Orbb veraw: © ‏اوه‎ Por of storage thot survives ‏و( له‎ © pproxtroted by cotctcttay wuliple copies ca dettact ‏ملحب جمدم‎ ‎Od. me ©Sbervehnts, Cork ced Cnakershe‏ ,تن 6 تیه 

صفحه 6:
وه ام( ‎Gicb's-Owraye‏ ۳۲ ‏مس لا موی خن ری لین نی(‎ separcie disks © copes con be ‏ام اد‎ sites ‏وه اوه مس وا‎ suck us Pre or © ‏وميك عملت‎ dota rousPer on sill resull to tervosistedt copies! Block trae Per ‏ذا انود ددم‎ ۰ ‏مه خی‎ ۱ © Dord Bake! ‏ال جر ایا مت‎ Bl Proeokn store weda Pro Poker dure doa ‏اس‎ (oe ooh): © Crete cp opera os Polos (asm unrkny LUD copes oP suck block): (0 Onte he kPorewatea ro he Pret phypird book. Okeo the Pirst wrte successPuly cowpletes, vars the score toPorcatiog wnt the seooed physicd block. he culpa cowpleted ‏واه رای‎ he seoved write success huly copes. @eedrwr Gyetre Oocowytr, Oe. me ©Sbervehnts, Cork ced Cnakershe 

صفحه 7:
+ | Ieopleweucios (Coc.) BH Protein store wed Brow Paki dure de traePer (oo.): BE Oops of a Hock ‏موه توت مد ان سدق رو‎ Po recover Prow Poker: 6 @irst Prod eeocwintedt bloke! ۱ ‏سوم( مشاه رو‎ he two copies of every disk bok. ‏ماه و‎ © ‏موی من مرن با موه لس‎ store (Dookie ROO pr ‏لخي همه اس‎ ۶ Ose the ‏ند سا امن رم مد وت‎ may be ‏ووه وس و ابو سب‎ bee, ۶ ‏میت لو‎ ٩۱۱0 ‏مود‎ 4P ether copy oP oe ‏مه ها وا لستجط ع ولا متسه‎ error (bad checker), venue thy the ober copy. “IP bots have a7 error, but ore dPPeredt, pverurie the seuoed black by the First block. سا0 لح 0 لا سواه 1 وجو ‎@eedrwr Gyetre Oocowytr, Oe.‏ 

صفحه 8:
Oxta Oovess © @hyetd books we those blocks residiry va the disk. ۲ uPPer blobs we the biochs residiey tewporady in wot wewory. © @lck woveweus between dish ood wot wewory we totfcted tus the ‏اطلام‎ uae ‏تصش‎ © tagn(®) those Pere the physiod block (Bt cocks ‏رم‎ © ‏)نحت‎ )©( irenePers the buPPer block (P to the disk, oad replaces the pproprinte physicd block there. ‎bool cope of ol cia‏ ات وا ‎hur ts private Wwork-ureu‏ ,7 ما ها ‎by ilove her.‏ لول ‎tews urreezed ond‏ ‎tow Xt oiled x,‏ مق و ‎P's bro copy of‏ © ‎© Oe weanve, Por skophoty, thot euch cote lew Pity to, ond is stored ‏رطس‎ ot stop blocks. ‎Od. ne ©Sbervehnts, Cork ced Cnakershe‏ ,تن 6 تیه 

صفحه 9:
+ Oxta Bovese (Ova) © Prnswivd troPers dota tews betweru sysiew bubPer blocks ood ts private ‏وود‎ ustey the Portes pperaiocs | © ret(0) seeks he uoke of skis few 20 the bool varkdble x, © vorte(X) cost the ude oF ‏جما‎ vartable 2, to dc tec OG i the bubPer book. © bok bese commends way vevessicie he tour of ‏مه تلا مه‎ ‏وا مخ ,مت سا موی‎ Bev whic ‏اتوم )ل‎ ts ot dread) fe wer. صم 1 31 تا سا ۲ رسیه ساحاند (00 )لجر ومرو تاوت © ‎Ol xbeequect accesses oe to the bod vopy.‏ © )ات تمه مت ره سا ‎Per‏ © 0 سس ۱ مات موه ‎Od. ne ©Sbervehnts, Cork ced Cnakershe‏ ,تن 6 تیه 

صفحه 10:
سس 6 اس ‎@ufFer‏ @OufFer Blk — 100 / 0 ۳ work aed ‏نام‎ 00 Gyetre Oocowytr, Oe. 

صفحه 11:
بزچموه) وه بجورمومو) سم بوب وه لب مس ‎he‏ بط مه فولات لد با بط ا بل ‎«pol‏ ۵ مج ‎ip‏ مه م۳ 30 مط ن 1 مسب لو ‎toh‏ و وا روا یت تایه بط سوم یلیر ‏له 0 ۰( ‎way be required Por 41١ (to vuiput @ ord‏ وه نوی لوق ‎have beeo wode but bePore oll oP thew‏ مشخ سا نس اه مج رون ‎we wade.‏ ‎ ‎4 ‏سا0 لح 0 لا سواه 1 مجو ‎Od.‏ ,تن 6 تیه 

صفحه 12:
+ Reever aud Giaicks (Cod) © ‏روت مت‎ despie Poles, we First cuiput iPorcatiod desorbiay the ‏ال سا رما نات مه ارو با سل‎ Be sity tw ‏اس‎ ‎© bxpbesed recovery, ord * ‏ومبسمسطصاد‎ ‎000 thot i, oo Pier the ober. سا0 لح 0 لا سواه 1 موه ‎@eedrwr Gyetre Oocowytr, Oe.‏ 

صفحه 13:
رو رمووو() لودع ظاعوورا B® by her oa okbe sine. © Phe loys a sequewe oF bey revords, ued cototctces ot revord oP update ‏وص انعد‎ va the ‏لول‎ © Ohkea inners 7, starts, tredsiers ‏رواخ‎ a <P, whales revord ۲ Before 1, exentes wrte(X), okey record <P, X, Oy O? 9 wien, where O, te he other oP XC bePore the wurie, ood Ot the uke ty be writes to X. © bet record antes tht 7) kor perPorwed a wrt ‏لک ما مد مه‎ her O, be ore the wrte, ond ull ave voke O,, Pier te wrte. Bl Whew 7) Brisker thet ottewed, be by record <P, ppew> t writes. Be coor Por cow tht bu records oe writed drecly tp okible stone (thot i, fey oe ot bP Pered) © Dw wercaches ‏ار‎ ‎© DePered database wodFivaion © Aexveddie dattbase wodPicdioa سا0 لح 0 لا سواه 1 موه ‎Od.‏ ,تن 6 تیه 

صفحه 14:
+ QsPerred Oxide DrdPodica ۲ ‏با‎ tePerred dotcbesr wodPivatve schewe records dl ‏روصا جما جا جمحش حالسب‎ but dePers dll the wrties ty oPter portal azat. © Ose nce he teeeuntiows ‏ره جیوه‎ ‎ke.‏ لو خی > مج برجا وه مه ‎B® wrte(X) perctica resus ta by reverd <7, X, O> betsy writen, where Ot he wow ihe Por X © ote! od uch ty oot ceeded Por this schewe ‎19 Dhe wrte cet perPorwed oa X ot his toe, bat ts dePerred. ‎Ohea 7, portly ores, <P poet? ‏ما سح‎ te kx ‎© Pind), the bby records ore read oad used ty oct) exerute the previously ‏لاو‎ ‎|۱۳ Gyetre Oocowytr, Oe. 7.0 ©Sbervehnts, Cork ced Cnakershe 

صفحه 15:
+ OsPerred Oxide DrdPodion (Orc) 18 7 ‏يحنى‎ recovery er ‏و‎ crash, 9 rowsuniiog oeeds ty be redooe P oad oaly P bot > ote orl? power oe tere ‏جا‎ he by. BE Redbkn o texecctoa 7, (redo?) sete he ude of ol cota few arched by te ‏ترجه با نا بو‎ vcher. اج مه بل ۲ عه ‎i examen the orice updos,‏ و با © مجحلا بيجا جا ‎kde recovery cote‏ ۶ B excep exeioes Ty ond T, (Ty ‏تسا فص‎ 1١ ‏:ل‎ Dy ress (P) Dy tread (OC) @:-@-80 0 0-400 Orte (®) wort (O) rewd () 6: ۵+0 ‎he‏ (0) هس ‏سا0 لح 0 لا سواه 1 موه ‎@eedrwr Gyetre Oocowytr, Oe.‏ 

صفحه 16:
+ OsPerred Oxide DrdPodion (Orc) © ew we show the bby os it ‏اه عم عدا اد سمه‎ <Tp start> <To start> <Tp start> <T, A, 950> ‏رن‎ ۸, 9502 >, 2, 950< >10, 8, 2050< ‏,و‎ 8, 2050< >, B, 2050> <Ty commit> >10 commit> <T, start> <T, start> <I, C, 600> = <T), C, 600> <T, commit> (b) (9) BPRixy oooh sinnp of eve oP orack is oe ki ose! (0) Oo rab wine weed w be her (b) redo ‏ام © ام( > هلچ سا شمه(‎ (0) vedo( Tg) ‏اسح‎ be perPorcoed Pobowed by rede T)) ‏اه‎ ‎<2, power nd <P, oouna> ore preset سا0 لح 0 لا سواه 1 موه ‎@eedrwr Gyetre Oocowytr, Oe.‏ 

صفحه 17:
+ towers Dokbwe DodPotva ۲ ۱ ‏ات مرن سل لت‎ dlows database updates oP ot oor ‏جد طلجت سا وا مس‎ the writes ure ‏لح‎ © swe uedoiey way be ceeded, upduie loys ust hove bois ok votue ced ceus uch 198 Opdate by record wast be writes before datas tec is usritest © Oe ws .nve thot the boy record is pulpal direcily tp stuble storage © Coa be extruded i poster by record pulp, 90 boon ‏جد‎ prior to exer ot of oc pulpa(®) operators Por a dota book @, of bt records correspon otros Bonet be Photed ty oktble ‏جمد‎ ‎BE Ouput of uechied books on tohe phire ooo koe bePore or Per ‏تمصي‎ ‏سیر‎ 18 OOrder ta whick blocks ure cual co be dP Perrot Brow the order ia ushick they: re writes, @eedrwr Gyetre Oocowytr, Oe. wae ©Sbervehnts, Cork ced Cnakershe 

صفحه 18:
ی + bow One Oud >, sho <P, B, WOO, SSO> 2, ۵, 6000, ۵6060 @=980 @=0090 >, wow > sta <0, 0, 20۵, 600< 0= 000 Dy ®, > ‏سوه‎ ‎4 ‎X.‏ ری انا سا ۵ اقا ‎|۱۳ Gyetre Oocowytr, Oe. ‏موه‎ 311 ‎ 

صفحه 19:
+ thowerkts Debwe DodPoctva (Ovu.) Bl Reve proche ‏ام موه ماب‎ of coe! © edb 1) restores the ‏سس‎ of ll dott tec upchaied by “D7 their ob chew, eee eee eee ee eet eee © reb(1) see he vcke of ol deta texee uted by Po the ew ches, yore Ponward Brow te Piet boo record Por 7, Bok opercioxe cam be kewpoied © Dro ‏حك رط‎ P he operdivg is executed wuliple fer the ePPevt is the socve a Pits exerted core * Deeded siwre operatives way yet re-exeruted dure ‏بمصحدم‎ ‎© hee reowvertay oer Pukire: ۱ ‏مس وا ۶ موی بانط( مس‎ he record ST, phat, bet does cet ovis he revord <P eouwA>. © Denton Poors be redoue Pe bx ovata bok the record <P) ett> arnt he record <7 pow. 19 )1( ‏له‎ opertiow oe perPorwed First, hea redy operciicas. ‎Od. «9.00 ©Sbervehnts, Cork ced Cnakershe‏ ,تن 6 تیه 

صفحه 20:
el ‏بت پا سل تیاه سب‎ tL oppeary ot threw Retaarey oP tern. <Ty start> <Ty start> <Ty Start> <Ty, A, 1000, 950> >10, ‏بق‎ 1000, 950< >10, A, 1000, 950> <To, B, 2000, 2050> >, 8, 2000, 2050< <Ty, B, 2000, 2050> <Ty commit> >10 commit> <T, start> <T; start> >11, C, 700, 600> — <T,, C, 700, 600> <T, commit> ‏م‎ (©) Cron wane Weak oer dae ae (0) canbe (1): © b revered » COOO ened Ov (OOO. (b) verbs (7) orn recy (Tp): Ow reotored POO, axed bea © ‏سف © لحت‎ ‏امهممسه 9060 لمی 060 باس‎ ‏طخ (م‎ )( cx re (P,): O axnd O wre ox ‏ای 990 و‎ 0 reopevand. Phew Oe ont» ODD هو ‎Od.‏ ,تن 6 تیه 

صفحه 21:
4 coated ۲ ‏لا‎ recovery proveder us deassed porter : ‏بط رام‎ pute bry fre oe ROAR ‏تاه سم ات مج ولج را موی له سب‎ ( ینجن ‏شا‎ updates to the database. © Greaves recovery procedure by periodicdly perPorwiay chevhpottay 1١ Ouputdl locprevords curredly resides io ait wewory poi stoble storage. Ouiput ol soodPied bubPer blocks te the disk. 2 Orie a log record > vhevkpotd> vain stable ‏يماد‎ سا0 لح 0 لا سواه 1 مه ‎Od.‏ ,تن 6 تیه 

صفحه 22:
+ Obsukporis (Orct.) © Outeg recovery we ceed tp ooosider oly the wost reved ‏شمه‎ ٩۳ ‏فكلا‎ ‎stated bePore the chechpotd, ‏اه اوه شحو لو‎ ۶ > امس مس جد لعو ابا اه له مس ‎(Goa backwards‏ ‎record‏ Ovniteur sconce backwards lo record <7 stat? ie Poured, ( ‏با وی بای لو(‎ port of ley Polowiesy cbove stad record. Gorter pend oF lpg can be tqeored durteg recovery, ced ooo be erased wheuever desired. 1) or dl ienexctvar (startar Pro Por hier) wih ‏صم‎ >1١ poet, ‏طحي ححص‎ )1١(, ))( ‏ادب سستجمها خم حصت جا بياج ج55‎ ( عص 41 مس اوه سس ‎Forward ict ihe boy, Por ol‏ رمموق) ‎hier wilt <7, vower>, exeruir redo(T).‏ سا0 لح 0 لا سواه 1 هو ‎Od.‏ ,تن 6 تیه 

صفحه 23:
‎Puke‏ سود امن ( اعسات صا سل جات ما تفت تاحه تلم لس ‎cos be‏ ۱ ‎Ty ew.‏ 110 9 علي ۱ ‏سا0 لح 0 لا سواه 1 و ‎Gyetre Oocowytr, Oe.‏ ۱۳| ‎ ‎ 

صفحه 24:
۳۹) سطله!۵) خا لحاس ها مت ع روم لها نا ماه معا بجوم و۵ ۲ ره اجه سا BE ‏وا بط مدا وج رد مت تم‎ of ‏مس و‎ the pared pe ‏تساه بط اج سل‎ pane ‏یف‎ ۲ ‏انا وج تمه با ی‎ to weve storage, suck trot state oP the database prior ‏لحم چا رو موه ماما و‎ ۱۳ wodPied durtoy exertion © Vo stat wth, bok the page tobles ore ideaicd. Ou curred paye tuble te used Por dat few unpesses dudey executiog oF the trocruriza. © Qkevever coy poe is cba ty be writes Por the Pirst tre © copy oF this pone te wade voip oo need poe. © Dhe correct par tuble te thea wade tp ‏تروصت ححا صا مادم‎ © Dhe update te perPorned oo the copy سا0 لح 0 لا سواه 1 هو ‎Od.‏ ,تن 6 تیه 

صفحه 25:
pigeon disk سا0 لح 0 لا سواه 1 ووم ‎Od.‏ ,تن 6 تیه 

صفحه 26:
+ 0 (Ghadow cad curred poop tobles oPier write i page shadow page table سا0 لح 0 لا سواه 1 هو ‎Od.‏ ,تن 6 تیه 

صفحه 27:
+ Obadow Parton (Oat) © Po cowl a ixxenios | عاص صا لوصوب مده ها جردم السب لد جاص اف .0 ©. 0 ‏سدم ححصت انحرف‎ ible to tots 6 0 he cures pax tbe he ce shank poe ble, oe Polos! © keep a pointer to the shadow pore table of o Pred (keowa) lpeaios vo deb. 6 6 22 12 22223 222 ‏د‎ fone etre rere ‏اك اك ل‎ ‎iy potuliy pannel puxe tobke na dks‏ جر ‏اوح مس رت ما سا ‎shadow pare‏ نا ‎Ocee porter‏ ۴ ‎© Wp reve & weeded Per 9 prick ‏,رهاط اجه موی مس تست‎ wien fhe shadowy pave tobe. ‎BE Pages owt ported iy Prow curreulshodbuy poy fable should be Breed (srarboxe ‏سیر‎ ‎Od. wor ©Sbervehnts, Cork ced Cnakershe‏ ,تن 6 تیه 

صفحه 28:
+ Chow Pater (Oow.) ۲ Odvoukes of shadou-pagry over berbused schewes © werhead oF writry log records نضا جا بمصصحص ‏ * 1 ‏:ص0‎ ‎۶ ‏و روت ال و له رون‎ » Coa be reckeed ‏يحاص بو‎ 3 poe thle pinwtwred the 0 O*-iree — Oo weed ty copy euire tee, ody weed tp copy poke ta the tree that bead to ‏سب ببس‎ © Cort overhead shiek even wit ‏مه و‎ + Deed ip Phok every urdhied pay, oud pace tobe © Dota ete Praxnoecied (rebied paves yet separced oa tsk) © ORer every resuniva cowplction, the dotabuse pages ooctatateg obd versizces oP swodhted data ceed to be garbage ‏لاس‎ © ler to exited deporte to olow trresurioes to rust cvarurreciy * Goster to extead log based schewes سا0 لح 0 لا سواه 1 هو ‎Od.‏ ,تن 6 تیه 

صفحه 29:
+ QRevovery Ok Cowured Traswirw ۴ De only te ‏نا وس وی وه و عالطا‎ 2222327 © 001 9 ‏وماد جاص‎ a sine dk bub Per ord a sted bot © 09 ‏تاها‎ bck oa have dots ters urchiied by mee or crore tracert BE Oe wm noe cowry roirol woken viret Luo-phare bhi © he the updkies oP veroercnited trrcsuntizas should oot be visible to ver ‏سور‎ ” Otkeruise how to perPorw verde PDC updates , thea DO updates (DB ‏را له ویو لو‎ TM bas to abort? ۱ ‏با ریسا‎ doce os desorbed eorter. ‏سا رو و وله اه سم رورا و‎ toterspersed ta the lou. ۲ ۱ ‏اد اه له سم پا‎ vo recovery have ty be chased © ste severd irneurives way be uive wheo a chevhpotd is perPorwed. سا0 لح 0 لا سواه 1 هو ‎@eedrwr Gyetre Oocowytr, Oe.‏ 

صفحه 30:
۲ ‏وه مان‎ perPorwed us bePore, except thot the chevhpoidt joy revord te ow of the Por < ‏حرا‎ ‎where ts the tet oP trocsunticos ucive of the tie of the cherkpotat ۶ ‏طولب موجه و(‎ or it prowess hte ‏واه‎ ie carted ot (ul roku tir kite) © Wheo he systew revwvers Prow a oresh, tf Pirst doer the Polowterg! ‏وی سا‎ ccd rede-let to ep (Goa the joy backwards Proc the pwd, stoppies uted the Pirst <chevkpotat L> vevord ts Pound. (Por ack record Pourd durtey the backward sro © Bike reper ‏دا‎ >1١ ‏<امجموو‎ , hl Te red-bet Phe records <7, phat, hea P Dye ‏,سحاد جامد‎ add ۰ ‏ال‎ ‎١ Por every Dial, B De ot i rect, ‏بل و لل‎ سا0 لح 0 لا سواه 1 ممم ‎@eedrwr Gyetre Oocowytr, Oe.‏ 

صفحه 31:
ths port vod coceiets oP recaps traneuntioes whick wast be untoce, ‏سوم سور لو‎ oF Finished trocsantiogs thot aust be redo. © Revvery ww oowinues os Pola: ‎bachivards Brow west reved record, stoppin whet‏ بجاو دحلم ‎eta records have bea earountened Por eveny 1١ ١‏ > ‎© Our he soon, perPors ede Por ‏د صا ما لا اس وا ات‎ trocsurtiog tr ucts. ‏الس را واه > فا ]مرا ‎bog Porwards Prow the <pheckporat L> record tke ead of the low.‏ 1۱ ‎© Orc he soon, perPorw rede Por cack oy record thot bel too ‏سل وه موم‎ ‎|۱۳ Gyetre Oocowytr, Oe. «7.00 ©Sbervehnts, Cork ced Cnakershe 

صفحه 32:
ی ۵ + وا مسا بط من میگ وضو عط ‎@o over he steps oP‏ 19 موه ,> <0 ,۵ ,0 ,> سس > و > <40 ,۵ ,۵ ,> | سا وه 4 مه دا موه ۱۶ و > ‎<P, 0, 0, 00<‏ <60 ,00 ,0 ,41> ‎٩ ۰(<‏ ممسواسای> بو > ‎<P, 0, 0, 60<‏ ‎<P, O, 0, 00>‏ سس > سا0 لح 0 لا سواه 1 هو ‎Od.‏ ,تن 6 تیه 

صفحه 33:
+ Log Record @uPPertc Bo bog revord bubPerted: oy records ore bubPered in watt wewory, fostead oP oP beta pulpal direnily to stable sore. © Lent records are pulpal ty stable pire hea a block of boc records ta ho babPer ts Pl, or a bog Poroe operon is executed. وممصم اه و رو ماه مه و او و ‎Low Borne‏ ۳ وه وه و( تمه ‎(eke he‏ © Geverd ley reverds con thus be cuipul ‏د عطاك‎ stage ‏مس شوه تون‎ the WO vost. سا0 لح 0 لا سواه 1 وو ‎Od.‏ ,تن 6 تیه 

صفحه 34:
+ Log Record BuPPertaq (Ovu.) BO ODhe nies below west be Polowed P log records ore bubPered: © beg reverts ore pulpal to stuble storage to the order ia whick they ore ‏لس‎ © ‏يجا جملا مت تسه ام با مه 1 مس‎ record > ‏ما مین‎ beew pu ty oti ‏واه‎ © ebore a block oF data te woke wewory ‏وه رال سا نا تن ع‎ records pertototey to dota fo trot blocks o7ust have bers cuigut to stable storage. ۱ ۶۳ nde ts dled the varte~dhead loygag or D@L rue ~ Girely speakie DOL ool requires ‏ای ص و موه طی‎ سا0 لح 0 لا سواه 1 موب ‎Od.‏ ,تن 6 تیه 

صفحه 35:
Octebose OuPPertay B Oottbase watdtoies oo ‏رو‎ bubPer of data beck © Okea a cew block is ceeded, P buPPer is Pull ow extstay blocks ueeds to be rewoved Prow buPPer 01P the block chose Por rewoud hes beeo updated, oust be muiput to dist 19 sa result oP the wortechead Irqgiay rule, Po block with veered updates is mulzut to disk, leg records wih uade toPoreaiiog Por the updates are pulpal te the bog oe stable store First. 1# Wow: ould be to progress ooo block wheo ite cuiput to disk. Ouc b ‏ابید‎ ‎us Polos. ۶ ‏مت سوه‎ o dott few, irrurion umquires exckisive lok oa block ovotaictcry the cote tec © bok co be releused pave the write te cowpleted. + ‏سا اس‎ held Por short duraica ore culled kaichew. © @ePore u block is vulput to disk, the sysiew urquires oo exchusive hatch oo the block * Cesures oo update coc be to progress 7 the blo سا0 لح 0 لا سواه 1 هو ‎Od.‏ ,تن 6 تیه 

صفحه 36:
+ OPPor Oucxpwed (Oow.) © Octubuse bubPer coc be topiewrcted ether © logo ed oF red wokrwewory reserved Por the dutdbuse, or © ‏ی مش‎ ۲ ‏وروی رو و تا با‎ her drawbacks: © Dewey ts portioued bePore-hoad between database buPPer ord ‏قاطا با ,امه‎ ۱۳۹ way chan, ood othougk opera, syste havws best hows swewory should be chivided up of ony tee, amet chooge the portiicatay oP wewory. سا0 لح 0 لا سواه 1 هو ‎@eedrwr Gyetre Oocowytr, Oe.‏ 

صفحه 37:
+ OPPor Oucxpwed (Oow.) ناه ره و ره لعج ‎eerily kop ented ta‏ بو ول له ۲ drawbacks: © Oheu operate syoew wer ‏ات وا‎ a pour hol ‏سل وا ,یی ما سم‎ epee Por coker pare, he poe ip unites tp swap space ou toh. © ‏سوا(‎ dotubwwe ‏یل‎ Wy wrte bPPer pave to ok, bPPer pose way be fa Suny space, ond way hve tobe red Brow swor spore oa doh ond ot to the ‏بط بحاصل من سول‎ ts extra VO! ‏مس‎ hed pacts problew. © ded) when swppin pul a cicbose bPPer pane, oper syste shod poss code bp dokbose, whick ia hua pups pace ‏واه تیه سل و‎ SUN ‏بط فده ط ید )ورد‎ Pars!) > Dad pari oon thus be worked, but exons operas ‏سوه‎ db ot روط امه وود ‎Od. wer ©Sbervehnts, Cork ced Cnakershe‏ ,تن 6 تیه 

صفحه 38:
+ (Pokee Wik bow oP Oodikdiy Gore ‎Powe waned w lose oP corvokile store‏ وق ا ‎Dechoique sive to chevhpoittey used to deul wits bss oP arcrupkatie storage‏ ۱ ‎Cericdicdly dcop the eure couiedt oF the chtcbose to stable strrace‏ © ‎© Oo resuriog wey be orive durioy the dup procedure; a provedure sitar to chevhpotatey wet tohe phase ‎> Quint dl os records murrediy reside ia waa weeny oxto sible sion * Output al buPPer blocks cote the cists. ١ Copy the ovotents oF the database to stable storace. ۱ ‏موجه سای من وان حول > ارو و تون‎ ©) ‏اد موم و‎ ۱ ‏مس ول مور‎ wost revect ducop. ۱ ‏مدع اه لاو لا عمجم اه مل له رها ع ون‎ ‎B® ‏وا للم و‎ olow trawsurtioes to be arive durioy ducrp; heowa os Pay dice or voor dope © Oil sity Paap checkportey ker ‏سا0 لح 0 لا سواه 1 هو ‎Od.‏ ,تن 6 تیه 

صفحه 39:


صفحه 40:
+ Gdanced Gecnery Techanee BH ‏و 60 بو ای سب امه ,مات رما روص و‎ oT ‏اش‎ ‎© Operations the @ ree tesertows und debtions ‏رات تا ما‎ © Dhey comun be ‏ول‎ by resioriogy okt voter (phyotrd vers), she pare « ook brewed, her rexeurines way have urduied the @tree. © Aeetenl, Keerivcs (resp. debetows) ure unror by ‏بمشججمت‎ 0 deleira (resp. feseriva) opercica (horus ‏مها چم‎ ued). Bor suck operon, ued ly records should ovotais he vedo operation ty be exerted © added gird uade bata ۰ ‏ات‎ ty phystod vans br. 15 Rede iPorwaina & byged phyeizdly (hol is, ‏سنج‎ voke Por euk write) eved Por suk operations © Ledted rede ‏یه امه بوچ‎ dotdbase stot oa desk way wt be ‏مه مرن‎ سا0 لح 0 لا سواه 1 موب ‎Od.‏ ,تن 6 تیه 

صفحه 41:
Operatica loguicy te door os Priaws: ‎stare, ba <7, O, operdtorbeda>. ere O, 16 3 wise‏ سین ما ‎idewPier of the operctios aki.‏ ‎Ohte ‏مه‎ tf exeruity, cord ley records wih physicd redo od physica sande ‏صمت جمدم ونام‎ booed. ‎Okeu pperdiva mops, <P, O, vperaorrad, D> is iced, where D ‏تمه له ی ماه مناج نا تج مهم مه تععین‎ ‎٩۴ ‏ورن تسوا موه بای‎ oops: ‎۱ ‏وه طجی و لح ها موق ‎ke physied vedls‏ © ‏ات رورت سل ولا 0000 1701000 ‎1 ‎© bxicd vad is perPorced ‏موق طلی لصوم :00 و‎ Por the persion ‏ای ها‎ Qedo of (Pier crust) stil uses ‎ ‎ ‎4 ‎@eedrwr Gyetre Oocowytr, Oe. ‏مهمه‎ 

صفحه 42:
+ Bdanued Qecciery Tectcique (od) ‎doar oe Poke‏ )1 ما جر اس( ‎BH Gow te by bahar‏ ‎Raby record <P, X, Oy On? & Poued, perPorw the vedo oad by‏ 0( ‎by record <7, X, OP‏ موسر اسرد ‎AB a <7, O, vperstowend, O> record te Pond‏ 60 مه ی ‎Rolback the operas brody using‏ * ‏اد ‎ane legged jet the‏ اس ام ومد لوط ‎Opies‏ ‏شحو ماه امه ‏و و روا اه ام ,سا مج اه امه ۲ 0۱) روص له ‎eed‏ ‏سوه 0 ,> و0 > ی اس وا بط و ‎beqo> is Pann‏ ‏سا0 لح 0 لا سواه 1 موب ‎@eedrwr Gyetre Oocowytr, Oe.‏ 

صفحه 43:
+ Bdanued Qecciery Tectcique (od) Bl Gown be by bukards (ov.): AR ‏لاصيا جبحا جا لمحي لاصلص و‎ ‏ممما جا لمعب <ا روطم سب ,0 ,11> د خاك ا‎ skip ‏لجح یط وا مس‎ >, 0, operctorbedts> © Pons. bop be soca when the revord <T, star? te Prvred Okla <P, bor? record ‏صما حا‎ boy Borne prt tr swe! © Ooses 9 and @ obove coo poor coy Phe ‏حول‎ orushes white a ‏ایا لام رما سا طسو‎ ۱ ‏و‎ oF ‏اوه سوه با‎ سا0 لح 0 لا سواه 1 موب ‎Gyetre Oocowytr, Oe.‏ ۱۳| 

صفحه 44:
+ ‏رس دس رگ‎ Teckel Ook) و موه و موم ات مه موه و۳ با اس را ماه < ‎(Goan fey Porward Brow ket‏ مه .1 Oreste oct veerot char he sia ces Prous را ول ۱ ای و للم جا ‎1١‏ لمتجنا جا ‎stan‏ > سا« ‎deleted Brow‏ 1 ,لوط ‎or <7, xbort>‏ حوموو > مسا( ‏لس ‎A his brags database to state os oP ‏وی جه أعنبد جه ای نت ی‎ ‏رنه شم‎ beeo redoce. ‎Wow ‏و نا مس موه لول‎ keeowpbte, tral tr, have cether ‏لسو يج رجور‎ wor bees Bully roled back. ‎@eedrwr Gyetre Oocowytr, Oe. wee ©Sbervehnts, Cork ced Cnakershe 

صفحه 45:
+ Bdanued Qecciery Tectcique (od) Recovery Brow systew oreek (ovct.) ©. Gra bu backwards, perPorcse wendy a ‏ام سس اه و‎ ‏ی‎ ‏وم و‎ are role back os desorbed eater. © hen <7, wa? ‏جا‎ Prue Por ‏موم و‎ 0 verte, wre a <P, ‏لمصصيت را و‎ ۶ ‏و > اب و بو‎ records hove beeo Pood Por of Pict ‏ال‎ 18/113957 adver the ef Penis oP ecowplete rocsuriioas (hose wih other powell le ey eer eee ere سا0 لح 0 لا سواه 1 موب ‎@eedrwr Gyetre Oocowytr, Oe.‏ 

صفحه 46:
+ Bdanued Qecciery Tectcique (od) Bl Okevhponatag i dice os Pokus: (0 0 ‏يجا اك تحف‎ records in ewer) 1 sible storage Out to doh dl waodPied bubPer books ١١ 0 ‏اتحف‎ bx oa otdble ‏باس را واه > و وت‎ ۱ oot dlowed to perPors coy onicos white chekpotiterg i it prowess. © Rugg chevhpototey dows trocsurtives i progress while the wost ive ‏واه خن وم روصت‎ one to progress ۱ va cen slide سا0 لح 0 لا سواه 1 موب ‎@eedrwr Gyetre Oocowytr, Oe.‏ 

صفحه 47:
+ Bdanued Qecciery Tectcique (od) ۲ ‏وه ریخ‎ is dour os Pols: ‏اس روا لآ و راو‎ Onte 3 <chevkpotet L> ley record ood Pore boy to stable storage ote tt D oP coodiPted buPPer blocks 0 Onw perni ‏اب موم وا شم‎ their orto Ovtput te cist ol soodPied buPPer blacks tet tet D blocks skoukl aot be updated ube beta ‏ان‎ ‏نطو‎ OL: ol ley records pertotatay ‏و‎ block wast be vutput bePore the block ts mutput Gtore 3 potter to the chevkpotat ‏من مانب ما ام و ام‎ hk © kes reomeriog 1 PRuzzy checkporal, stot sca Brow the chevkpotd record ported io by ‏ماس‎ © bow records bePore et_chevkpord hove their upduies reAlevied tr dutcbase pods, cod ceed oot be redo. ‎eens checkporis, where sysiew hed crashed white perPorcicny‏ و ‎Ez‏ رای لا ‎checkpotat, are‏ ‎@eedrwr Gyetre Oocowytr, Oe. ner ©Sbervehnts, Cork ced Cnakershe ‎ 

صفحه 48:
@OR166 Revovery Okprikw 

صفحه 49:
+ 2۳۷0۵ ۲ 0186 ‏و و‎ oF he ot recovery wrth ‎cuvercus upitvizaives to reduce overheads durtay oral‏ وا و ‎provessrn| ued ip speed up repovery‏ ‎© Dhe “odveued recovery okprihew” we sted carder tr wrdeled Pier ORIEG, but gredly ‏اه وم توا موه‎ ‏0 مد سم اه 0 ۲ سس نا رل نا (00) امه مصصوو بها تسوا ۱ ‏سا راما عم ‎whol updates‏ رام و همم و و()۵)را موی ۱ اد و ‎pphed to‏ ‎٠١ ‏لصم ادصحاصصيوا)‎ Oiny poe toble ty avord umevessary redos durtag recovery ‎(0) Cheap checkpotatey that ody records iePorwatica about ddy payee, ord does ot require diy poops te be vorites cut ot chechpotc Ee ‎* Qore cowie up vo cork of the ubove ... ‏سا0 لح 0 لا سواه 1 موب ‎Od.‏ ,تن 6 تیه 

صفحه 50:
@OR166 Optcazatows © ‏یاه‎ re © OR Revted poe is physiol) deuiPi|d, uciod wikia page coc be boyd © Osed to reduce looper, pvereerds ‎a repo is deleted ced dl ther records hove ty be woved‏ ات بو ‎to Pil kote‏ ‎« ‏جوا‎ rede oa log feet the record delet ‎> Physicdd rede would require lrgsiog oP ob cod ce vokes Por swuck oF the pace ‏ب دستص جف عاك جا اتحودج ‎Requires pace ty be‏ ‎hardware (ROD, dso supported by sve dst:‏ ارت مه و بوچ مود ‎~— eexeplete page cuiput coc be deterted by checks ‏رجا‎ ‎> Out extra ‏ات‎ ore required Por recovery > Drected as ‏ه‎ edt ‏لت‎ ‎Od.‏ ,تن 6 تیه 

صفحه 51:
+ ORWO Ore Orwiures Bboy seqewe wevber (UG6O) dealer euch by revord © ‏اص(‎ be sequecidly rocreasicry © Dypiedy oa of Peet Prow beara oF boy Ble ty olow Past aces ‏ما را ری مامت رات‎ © Gok pe ‏ی‎ o Pagel GO whick is the LGO of the lest ley record whose ‏ام‎ ore reReried oo the poe © Dodie 3 page ۱ ‏ام(‎ the por, cod write the by repord * Opdkie the pace ۱ Revord the LGO of the boy revord it PaeLOO ۱ ‏عم سل(‎ © Pore Phish to disk O-hiches poe » De poe stite oo desk & opercica occa! ~ Required ip support physiplogical recy © ‏توص يتيك لحك و رام‎ ty preved repeded redo Pe bs سا0 لح 0 لا سواه 1 مهمه ‎Od.‏ ,تن 6 تیه 

صفحه 52:
+ @RWO Ove Orrwiurse (Ou) © Goth by record crates LEO oP previous log record oP the ‏او وی‎ LSN TransId_PrevLSN_ RedoInfo UndoInfo be thrtog record way be Kopio4 ‎vowpeuratoa boy revord (OLR) werd ty bt‏ اه او وا رال تس تا ‎dureg recovery thot cever ceed to be ure‏ لا شم ‎© Ob gene the roe of opercioc-ubont bby reverds used in odvaued revovery ‏ماه ‎© ‏مرا‎ a Pek! OodoOex GO to wie ext (carter) record ty be uedooe ۱ Records to betwero woud hove dlready beed vedo * Required to avoid repected ade oP ‏ات وی امه‎ ‎LGO PraclO® ‏تایه وولو‎ ‎ ‎ ‎@eedrwr Gyetre Oocowytr, Oe. ‎ ‎ ‎ ‎ ‎ 

صفحه 53:
+ @RWO Ove Orrwiurse (Ou) B On@up hdr © bet oP paves in the buPPer thot hove bers updated ‏اس ۳ ,مین و‎ suck poe ١ PagelhOO of the poe © RebGO & wa bGO suck that ley records bePore thie LOO ‏راد عم‎ ‏مسا‎ uppled to he page versioa po desk ~ Getto cored ead of log used 0 poor is oserted tate dirty poe table (hot bePore bets uechied) ~ Recorded ic chevkpoints, helps to einioie redo work: © Obechport by record ۰ ‏عون‎ ‎۱ ‏تم سوت تن با لجی و۳۳‎ ۱ ۳ ‏سره وی‎ rocsuntion, Lost GO, the LGD oF the hast logy record ‏امد‎ by the traceractin © Cee positon va disk wies LEO of hast excpteted سا0 لح 0 لا سواه 1 وو ‎@eedrwr Gyetre Oocowytr, Oe.‏ 

صفحه 54:
+ @RWO Revowvery Okoritw ORCC recovery woken hree pwses Bl ‏ةا‎ pose! Orterwkes © hick ‏و مس‎ © Whick pee were diy (deh versio oot up to che) of eve oP rch © Redd GO: LEO Brow whick rede shod stort B Reb pest © Repeus Keto, ‏مه گم‎ Red GO » Rech GO oad Parpb GOs oe weed to avoid redo wots dread phere oe ier © Ocdo pes: © Robs back ol eer p te trocsurtions ١ ‏موه‎ whose ober wos coxplete carer are ot urdooe وت عیشت ول وی تشم وا وی وا له وم تم ‎(Key‏ © ‎oe required‏ لو و له لها سا0 لح 0 لا سواه 1 موب ‎Od.‏ ,تن 6 تیه 

صفحه 55:
+ ORWO Revovery: Ordos عم سای ‎log record‏ اسان وی با ما وت ۲ ‎Proce beg record‏ ویو( نا لت © حاحب اسه ناك ‎Gets Redo GO = wit oP RecLOOs of ofl pres is Oi‏ © ‎ww paves ore drip, Redo GO = chevhport record's LEO‏ و و ‎١‏ ‎Gets verdo-tet = bet oF trocsurtivas to chevhpotat ly record‏ © © Reus LG of ket by reverd ‏بویت و مس ات و‎ hevhpotad ley record Gone Porwerd Prow checkpoint © Ouest page... سا0 لح 0 لا سواه 1 هو ‎@eedrwr Gyetre Oocowytr, Oe.‏ 

صفحه 56:
+ @RWO Revvery: Oude (Ova) ude poe (ov.) © Goo Porword Prow chevhpornt © 4AP cay ley record Pourd Por tronsuricd oot fo uerdo-tet, odds troerantia to vedo tet © Ohevever oo updaie bby revord is Pound "AP pose is oot to Oi Pace Dube, itis ockded with Rec GO set to LEO of the update logy record © 1Piecesuniiva ead Ioy record Pood, delete ierosurica Proc ucrdo-tst © (leeps track of fest log ‏ای و مس ات و لس‎ * Day be ceeded Por hater verde ‎of onli puss!‏ لا ‎© Ret GO deterwices where ty stort neds pose ‎© Rech GO Por cack pore tr Or APagePable used ‏ی و‎ redo worl: ‎© 31 ‏له لیا مومسم‎ to be roled back ‏سا0 لح 0 لا سواه 1 هو ‎@eedrwr Gyetre Oocowytr, Oe.‏ 

صفحه 57:
+ ORWO Reb Pow Qedy Poss: Repeus history by reployioy every aniiva oot dread rePievied te the page oa disk, oe Pols: © Goere Ponwerd Prow Redo GO. Ohevever oo update boy revord te Pouerd: (AP tke poe te ont it Dit <Page Mable or the LOGO of the bog record is bees thao the Rec) GO oP the pare io Dit APaceDable, thea ship the boy record » Olkenwiee Petck the page Pro dick. IP the PagelGO of the poe Petched Proc disk ts fess thoc the LOGO of the logy record, rede the boxy record WOTE: P ether test is ceguive the ePPevts oP the bey record ‏راد عم‎ sippeured vo the page. Pirst test avoids evec Petchioy the poce Prow dish! ‎Od. wer ©Sbervehnts, Cork ced Cnakershe‏ ,تن 6 تیه 

صفحه 58:
+ @RWO Oud Ore Whe ‏رمسو جم صا طحب من‎ Por ot uzdate boy record © @evercte ¢ OUR ooctotatey ter vedo orton perPorued (aciocy perPorcned dura ucds are ۳ ‏.(ب فح‎ ١ ‏ساسا ین و بت لو لس بان‎ © Get DadDex OO of the OUR to the PreXOO ude oP te update oy reverd © Orrows techate DerdoDend OO vale BORGO apport partd rolback © Oeed &.q. to hood: deudooks by roltey back Ket rou to rekecse read. looker © Prqure techoates Porward ovtivey Pier portal roaches * reverds 9 aed & rata, hater S oad O, thea Pull rollback: 0-0 -) oe ‘ebb 06% ۰ سا0 لح 0 لا سواه 1 هو ‎Od.‏ ,تن 6 تیه 

صفحه 59:
+ ORWO: Oud Poe Onde pes © Cerforws bochwed soo oo oy vedoiry ol trresurtizg io verdo-tst ۶ ‏و نله‎ ppitetzed by shipptey ueerded ley records os Pols? © Dext LGO te be vedere Por eack trocsoriva set to LOO of fest log record Por trassortiog Pourd by ocdleis poss. * @teuck step pick haryest oF these LGOs to undo, ship back tt ood ede ۱ )۳ ‏اس روا ه بل‎ © Cor ordeary ley records, set cent bGO to be uedoor Por ‏وا ما‎ Pred ‏اس با سا وا لت‎ - ۳ ‏وا موی‎ records (OL(Rs) set cent LGO te be vende ‏م‎ ‎OcrkDen SO crted fa the bog record » Ol ictervediay records are shipped siooe they would ove beec urd ‏باه‎ ای لام ‎perPorwed‏ ول( ۲ سا0 لح 0 لا سواه 1 ووم ‎Od.‏ ,تن 6 تیه 

صفحه 60:
+ 0 R106 ‏سیم‎ © Revery Ietependewe © axes can be recovered indepeadeniy oP vers » ‏سره ۴ وج‎ deb pours Pal hey oon be recovered Brow a backup while cher ‏را وم‎ werd ۲ Gaweponis: 5 10 ‏سامحم‎ coo revo savepoidis aad rol back to a suvepoiat * OsePul Por copter trassurios © @bp wed to rolbook just eo ugk tp reteuse bck ‏ام من‎ سا0 لح 0 لا سواه 1 هو ‎@eedrwr Gyetre Oocowytr, Oe.‏ 

صفحه 61:
+ ‏ان‎ ۵0۷۵6 Prawres (Oru) © re-qraiced lochicry: ۱ thal peril tuple level lochieg vo tedices cart be ‏لوص‎ لسعمدصعه متكت ,على امحتصيوام مها ,صاصر يحل ‎require loyical‏ ۳ ۱ سم ا 13 ‏وی ۳ هو رم‎ © Ory pore tbl coo be used to prePetch payer duro redo © ‏رین‎ order redo is possible: * reds oan be posippaed ooo page betoy Peicked Prow dk, ced ‏وم ات لاو‎ if Petched. + ‏موی طسو رها اه طایمه()‎ poutnue ty be processed |۱۳ Gyetre Oocowytr, Oe. «7.00 ©Sbervehnts, Cork ced Cnakershe 

صفحه 62:


صفحه 63:
م6 وله نس + ۲ Rewnte backup systews provide high wwvulubliy by olowiey ieneuniva processtay 17 ‏مجك صوقموم‎ P the priory ste ts destroyed. network log records @eedrwr Gyetre Oocowytr, Oe. ee ©Sbervehnts, Cork ced Cnakershe 

صفحه 64:
+ Qewoe Backup Oystews (Ovd.) © Oetriva of ature: Buchup site cust detect wheo pricey site ‏اه عم‎ © to deteenick primary ste Poker Prow bol ‏اوه وی سل‎ po ‏دادج فى‎ hake beter the privcary ocd the newts backup. © Procter of wait © Dp tohe over ood backup ste Pst perPorn revovery ‏جل جاص‎ copy oP the database ced ol the brag records thos reveived Prow the priory. و مسا مومس له سل ‎ore‏ ما لاو ‎Dos,‏ ‏اس لا ©) Okeu the backup site tches over processiay it beowes the oew priory © Do trxePer ovine buck to old priwary when tl recovers, ol privary wet reveive rede loye Prow the old backup ood apply ol upchites locally. سا0 لح 0 لا سواه 1 وج ‎Od.‏ ,تن 6 تیه 

صفحه 65:
+ QRewoe Backup Opsiews (Ovw.) ۲ ‏تومیر ن و‎ Vo reduce dehy in tohevver, backup ste pericdicaly proceses the redy bog revords (in eP Pen, perPorwieg recovery Prow previous database tute), perPores o chevhpoid, ood coo thea delete parker ports oP the oy. Bt Opa ‏موه‎ pers very Post hover: © @ochw ooonily processes redb ley record os hey onrive, applic be updates torch. © Whe Poke of the primary & deterted the backup rolls back ‏امس‎ ‏و له ,مات‎ ready ‏امس تن و و‎ 18 Qlercaive to rewrite buckup! dstributed databuse wih rephoated cata © Rewote buckup is Poster oad cheuper, but less toleroat to Pothure ۱ ‏امه جص‎ ۰ )( 06 سا0 لح 0 لا سواه 1 ووه ‎Gyetre Oocowytr, Oe.‏ ۱۳| 

صفحه 66:
+ QReuvte wh Opstews (Ovu.) BE Crewe darby of urckies by delavieny irmeanioa coco uci uci ts bored ot backup} werd ‏تال خن سول تن وحم روا روا‎ ۲ ‏ها ع لس ایحا و موه وه او ییون‎ primary © Problew: urckies way wl arive of buckup bePore it tokes over. Duende! peer wheo trocsuntiod's peer fry nepord is varie ot prany cond backup © Reckees wokbliy ‏لي وه‎ ether site Paks. ۲ Dives: proceed op in hue-very-de B bots primary ond backup ore orive. ۴ waly the primary fe ‏لس وا ی جح مرو و رن ما با رهش‎ ‏مر اب من‎ © Ceter wulbiliy thon bor-veryrde; woke problew of bret trowsartiocr i 7 اه سا0 لح 0 لا سواه 1 هو ‎Gyetre Oocowytr, Oe.‏ ۱۳| 

صفحه 67:
Gad oP Okaper 

صفحه 68:
input(A) output(B) هم 00 6 ‏,تن‎ Od. 

صفحه 69:
Corton oP te Odubwse boy Oorrespouiag to To aod % <T start> <To, A, 950> >10, 8, 2050< <أأماكامه 10> <T, start> <T,, C, 600> <T, commit> + @eedrwr Gyetre Oocowytr, Oe. 

صفحه 70:
۹ OP te bog wad Outebase Oorrespoadiny to T, aod % Log Database <Ty start> <Ty, A, 950> <Ty, B, 2050> <Ty commit> <T, start> <T,, C, 600> <T, commit> سا0 لح 0 لا سواه 1 هجوج ‎@eedrwr Gyetre Oocowytr, Oe.‏ 

صفحه 71:
‎Tr wh T,‏ پاچ یم ‎<Ty start> <Tp, A, 1000, 950> <Tp, B, 2000, 2050> ‎<Ty commit> ‎<T, start> ‎<T,, C, 700, 600> <T, commit> ‎ ‎@eedrwr Gyetre Oocowytr, Oe. wee ©Sbervehnts, Cork ced Cnakershe 

صفحه 72:
Grae oP Gystew bog ard Oatbose Oorrespradiay tio Ty wl T, Log Database <Tp start> >10, ‏بك‎ 1000, 950> >10, B, 2000, 2050> >10 commit> >11 start> <T,, C, 700, 600> <T, commit> @eedrwr Gyetre Oocowytr, Oe. 

Chapter 17: Recovery System Database System Concepts ©Silberschatz, Korth and Sudarshan See www.db-book.com for conditions on re-use Chapter 17: Recovery System  Failure Classification  Storage Structure  Recovery and Atomicity  Log-Based Recovery  Shadow Paging  Recovery With Concurrent Transactions  Buffer Management  Failure with Loss of Nonvolatile Storage  Advanced Recovery Techniques  ARIES Recovery Algorithm  Remote Backup Systems Database System Concepts, 5th Ed. 17.2 ©Silberschatz, Korth and Sudarshan Failure Classification   Transaction failure :  Logical errors: transaction cannot complete due to some internal error condition  System errors: the database system must terminate an active transaction due to an error condition (e.g., deadlock) System crash: a power failure or other hardware or software failure causes the system to crash.  Fail-stop assumption: non-volatile storage contents are assumed to not be corrupted by system crash   Database systems have numerous integrity checks to prevent corruption of disk data Disk failure: a head crash or similar disk failure destroys all or part of disk storage  Destruction is assumed to be detectable: disk drives use checksums to detect failures Database System Concepts, 5th Ed. 17.3 ©Silberschatz, Korth and Sudarshan Recovery Algorithms  Recovery algorithms are techniques to ensure database consistency and transaction atomicity and durability despite failures   Focus of this chapter Recovery algorithms have two parts 1. Actions taken during normal transaction processing to ensure enough information exists to recover from failures 2. Actions taken after a failure to recover the database contents to a state that ensures atomicity, consistency and durability Database System Concepts, 5th Ed. 17.4 ©Silberschatz, Korth and Sudarshan Storage Structure    Volatile storage:  does not survive system crashes  examples: main memory, cache memory Nonvolatile storage:  survives system crashes  examples: disk, tape, flash memory, non-volatile (battery backed up) RAM Stable storage:  a mythical form of storage that survives all failures  approximated by maintaining multiple copies on distinct nonvolatile media Database System Concepts, 5th Ed. 17.5 ©Silberschatz, Korth and Sudarshan Stable-Storage Implementation  Maintain multiple copies of each block on separate disks    copies can be at remote sites to protect against disasters such as fire or flooding. Failure during data transfer can still result in inconsistent copies: Block transfer can result in  Successful completion  Partial failure: destination block has incorrect information  Total failure: destination block was never updated Protecting storage media from failure during data transfer (one solution):  Execute output operation as follows (assuming two copies of each block): 1. Write the information onto the first physical block. 2. When the first write successfully completes, write the same information onto the second physical block. 3. The output is completed only after the second write successfully completes. Database System Concepts, 5th Ed. 17.6 ©Silberschatz, Korth and Sudarshan Stable-Storage Implementation (Cont.)  Protecting storage media from failure during data transfer (cont.):  Copies of a block may differ due to failure during output operation. To recover from failure: 1. 2. First find inconsistent blocks: 1. Expensive solution: Compare the two copies of every disk block. 2. Better solution:  Record in-progress disk writes on non-volatile storage (Non-volatile RAM or special area of disk).  Use this information during recovery to find blocks that may be inconsistent, and only compare copies of these.  Used in hardware RAID systems If either copy of an inconsistent block is detected to have an error (bad checksum), overwrite it by the other copy. If both have no error, but are different, overwrite the second block by the first block. Database System Concepts, 5th Ed. 17.7 ©Silberschatz, Korth and Sudarshan Data Access  Physical blocks are those blocks residing on the disk.  Buffer blocks are the blocks residing temporarily in main memory.  Block movements between disk and main memory are initiated through the following two operations:   input(B) transfers the physical block B to main memory.  output(B) transfers the buffer block B to the disk, and replaces the appropriate physical block there. Each transaction Ti has its private work-area in which local copies of all data items accessed and updated by it are kept.   Ti's local copy of a data item X is called xi. We assume, for simplicity, that each data item fits in, and is stored inside, a single block. Database System Concepts, 5th Ed. 17.8 ©Silberschatz, Korth and Sudarshan Data Access (Cont.)    Transaction transfers data items between system buffer blocks and its private work-area using the following operations :  read(X) assigns the value of data item X to the local variable xi.  write(X) assigns the value of local variable xi to data item {X} in the buffer block.  both these commands may necessitate the issue of an input(BX) instruction before the assignment, if the block BX in which X resides is not already in memory. Transactions  Perform read(X) while accessing X for the first time;  All subsequent accesses are to the local copy.  After last access, transaction executes write(X). output(BX) need not immediately follow write(X). System can perform the output operation when it deems fit. Database System Concepts, 5th Ed. 17.9 ©Silberschatz, Korth and Sudarshan Example of Data Access buffer input(A) Buffer Block A x Buffer Block B Y A output(B) read(X) write(Y) x2 x1 B disk y1 work area of T1 work area of T2 memory Database System Concepts, 5th Ed. 17.10 ©Silberschatz, Korth and Sudarshan Recovery and Atomicity  Modifying the database without ensuring that the transaction will commit may leave the database in an inconsistent state.  Consider transaction Ti that transfers $50 from account A to account B; goal is either to perform all database modifications made by Ti or none at all.  Several output operations may be required for Ti (to output A and B). A failure may occur after one of these modifications have been made but before all of them are made. Database System Concepts, 5th Ed. 17.11 ©Silberschatz, Korth and Sudarshan Recovery and Atomicity (Cont.)  To ensure atomicity despite failures, we first output information describing the modifications to stable storage without modifying the database itself.  We study two approaches:   log-based recovery, and  shadow-paging We assume (initially) that transactions run serially, that is, one after the other. Database System Concepts, 5th Ed. 17.12 ©Silberschatz, Korth and Sudarshan Log-Based Recovery  A log is kept on stable storage.  The log is a sequence of log records, and maintains a record of update activities on the database.  When transaction Ti starts, it registers itself by writing a <Ti start>log record  Before Ti executes write(X), a log record <Ti, X, V1, V2> is written, where V1 is the value of X before the write, and V2 is the value to be written to X.  Log record notes that Ti has performed a write on data item Xj Xj had value V1 before the write, and will have value V2 after the write.  When Ti finishes it last statement, the log record <Ti commit> is written.  We assume for now that log records are written directly to stable storage (that is, they are not buffered)  Two approaches using logs  Deferred database modification  Immediate database modification Database System Concepts, 5th Ed. 17.13 ©Silberschatz, Korth and Sudarshan Deferred Database Modification  The deferred database modification scheme records all modifications to the log, but defers all the writes to after partial commit.  Assume that transactions execute serially  Transaction starts by writing <Ti start> record to log.  A write(X) operation results in a log record <Ti, X, V> being written, where V is the new value for X  Note: old value is not needed for this scheme  The write is not performed on X at this time, but is deferred.  When Ti partially commits, <Ti commit> is written to the log  Finally, the log records are read and used to actually execute the previously deferred writes. Database System Concepts, 5th Ed. 17.14 ©Silberschatz, Korth and Sudarshan Deferred Database Modification (Cont.)  During recovery after a crash, a transaction needs to be redone if and only if both <Ti start> and<Ti commit> are there in the log.  Redoing a transaction Ti ( redoTi) sets the value of all data items updated by the transaction to the new values.  Crashes can occur while   the transaction is executing the original updates, or  while recovery action is being taken example transactions T0 and T1 (T0 executes before T1): T0: read (A) T1 : read (C) A: - A - 50 C:- C- 100 Write (A) write (C) read (B) B:- B + 50 write (B) Database System Concepts, 5th Ed. 17.15 ©Silberschatz, Korth and Sudarshan Deferred Database Modification (Cont.)  Below we show the log as it appears at three instances of time.  If log on stable storage at time of crash is as in case: (a) No redo actions need to be taken (b) redo(T0) must be performed since <T0 commit> is present (c) redo(T0) must be performed followed by redo(T1) since <T0 commit> and <Ti commit> are present Database System Concepts, 5th Ed. 17.16 ©Silberschatz, Korth and Sudarshan Immediate Database Modification  The immediate database modification scheme allows database updates of an uncommitted transaction to be made as the writes are issued   since undoing may be needed, update logs must have both old value and new value Update log record must be written before database item is written  We assume that the log record is output directly to stable storage  Can be extended to postpone log record output, so long as prior to execution of an output(B) operation for a data block B, all log records corresponding to items B must be flushed to stable storage  Output of updated blocks can take place at any time before or after transaction commit  Order in which blocks are output can be different from the order in which they are written. Database System Concepts, 5th Ed. 17.17 ©Silberschatz, Korth and Sudarshan Immediate Database Modification Example Log Write Output <T0 start> <T0, A, 1000, 950> To, B, 2000, 2050 A = 950 B = 2050 <T0 commit> <T1 start> x1 <T1, C, 700, 600> C = 600 BB, BC <T1 commit> BA  Note: BX denotes block containing X. Database System Concepts, 5th Ed. 17.18 ©Silberschatz, Korth and Sudarshan Immediate Database Modification (Cont.)   Recovery procedure has two operations instead of one:  undo(Ti) restores the value of all data items updated by Ti to their old values, going backwards from the last log record for Ti  redo(Ti) sets the value of all data items updated by Ti to the new values, going forward from the first log record for Ti Both operations must be idempotent  That is, even if the operation is executed multiple times the effect is the same as if it is executed once    Needed since operations may get re-executed during recovery When recovering after failure:  Transaction Ti needs to be undone if the log contains the record <Ti start>, but does not contain the record <Ti commit>.  Transaction Ti needs to be redone if the log contains both the record <Ti start> and the record <Ti commit>. Undo operations are performed first, then redo operations. Database System Concepts, 5th Ed. 17.19 ©Silberschatz, Korth and Sudarshan Immediate DB Modification Recovery Example Below we show the log as it appears at three instances of time. Recovery actions in each case above are: (a) undo (T0): B is restored to 2000 and A to 1000. (b) undo (T1) and redo (T0): C is restored to 700, and then A and B are set to 950 and 2050 respectively. (c) redo (T0) and redo (T1): A and B are set to 950 and 2050 respectively. Then C is set to 600 Database System Concepts, 5th Ed. 17.20 ©Silberschatz, Korth and Sudarshan Checkpoints   Problems in recovery procedure as discussed earlier : 1. searching the entire log is time-consuming 2. we might unnecessarily redo transactions which have already 3. output their updates to the database. Streamline recovery procedure by periodically performing checkpointing 1. Output all log records currently residing in main memory onto stable storage. 2. Output all modified buffer blocks to the disk. 3. Write a log record < checkpoint> onto stable storage. Database System Concepts, 5th Ed. 17.21 ©Silberschatz, Korth and Sudarshan Checkpoints (Cont.)  During recovery we need to consider only the most recent transaction T i that started before the checkpoint, and transactions that started after Ti. 1. Scan backwards from end of log to find the most recent <checkpoint> record 2. Continue scanning backwards till a record <Ti start> is found. 3. Need only consider the part of log following above start record. Earlier part of log can be ignored during recovery, and can be erased whenever desired. 4. For all transactions (starting from Ti or later) with no <Ti commit>, execute undo(Ti). (Done only in case of immediate modification.) 5. Scanning forward in the log, for all transactions starting later with a <Ti commit>, execute redo(Ti). Database System Concepts, 5th Ed. 17.22 from Ti or ©Silberschatz, Korth and Sudarshan Example of Checkpoints Tf Tc T1 T2 T3 T4 system failure checkpoint  T1 can be ignored (updates already output to disk due to checkpoint)  T2 and T3 redone.  T4 undone Database System Concepts, 5th Ed. 17.23 ©Silberschatz, Korth and Sudarshan Shadow Paging  Shadow paging is an alternative to log-based recovery; this scheme is useful if transactions execute serially  Idea: maintain two page tables during the lifetime of a transaction –the current page table, and the shadow page table  Store the shadow page table in nonvolatile storage, such that state of the database prior to transaction execution may be recovered.  Shadow page table is never modified during execution  To start with, both the page tables are identical. Only current page table is used for data item accesses during execution of the transaction.  Whenever any page is about to be written for the first time  A copy of this page is made onto an unused page.  The current page table is then made to point to the copy  The update is performed on the copy Database System Concepts, 5th Ed. 17.24 ©Silberschatz, Korth and Sudarshan Sample Page Table Database System Concepts, 5th Ed. 17.25 ©Silberschatz, Korth and Sudarshan Example of Shadow Paging Shadow and current page tables after write to page 4 Database System Concepts, 5th Ed. 17.26 ©Silberschatz, Korth and Sudarshan Shadow Paging (Cont.)  To commit a transaction : 1. Flush all modified pages in main memory to disk 2. Output current page table to disk 3. Make the current page table the new shadow page table, as follows:  keep a pointer to the shadow page table at a fixed (known) location on disk.  to make the current page table the new shadow page table, simply update the pointer to point to current page table on disk  Once pointer to shadow page table has been written, transaction is committed.  No recovery is needed after a crash — new transactions can start right away, using the shadow page table.  Pages not pointed to from current/shadow page table should be freed (garbage collected). Database System Concepts, 5th Ed. 17.27 ©Silberschatz, Korth and Sudarshan Show Paging (Cont.)   Advantages of shadow-paging over log-based schemes  no overhead of writing log records  recovery is trivial Disadvantages :  Copying the entire page table is very expensive   Can be reduced by using a page table structured like a B +-tree – No need to copy entire tree, only need to copy paths in the tree that lead to updated leaf nodes Commit overhead is high even with above extension  Need to flush every updated page, and page table  Data gets fragmented (related pages get separated on disk)  After every transaction completion, the database pages containing old versions of modified data need to be garbage collected  Hard to extend algorithm to allow transactions to run concurrently  Easier to extend log based schemes Database System Concepts, 5th Ed. 17.28 ©Silberschatz, Korth and Sudarshan Recovery With Concurrent Transactions   We modify the log-based recovery schemes to allow multiple transactions to execute concurrently.  All transactions share a single disk buffer and a single log  A buffer block can have data items updated by one or more transactions We assume concurrency control using strict two-phase locking;  i.e. the updates of uncommitted transactions should not be visible to other transactions   Logging is done as described earlier.   Otherwise how to perform undo if T1 updates A, then T2 updates A and commits, and finally T1 has to abort? Log records of different transactions may be interspersed in the log. The checkpointing technique and actions taken on recovery have to be changed  since several transactions may be active when a checkpoint is performed. Database System Concepts, 5th Ed. 17.29 ©Silberschatz, Korth and Sudarshan Recovery With Concurrent Transactions (Cont.)  Checkpoints are performed as before, except that the checkpoint log record is now of the form < checkpoint L> where L is the list of transactions active at the time of the checkpoint   We assume no updates are in progress while the checkpoint is carried out (will relax this later) When the system recovers from a crash, it first does the following: 1. Initialize undo-list and redo-list to empty 2. Scan the log backwards from the end, stopping when the first <checkpoint L> record is found. For each record found during the backward scan: 3.  if the record is <Ti commit>, add Ti to redo-list  if the record is <Ti start>, then if Ti is not in redo-list, add Ti to undo-list For every Ti in L, if Ti is not in redo-list, add Ti to undo-list Database System Concepts, 5th Ed. 17.30 ©Silberschatz, Korth and Sudarshan Recovery With Concurrent Transactions (Cont.)  At this point undo-list consists of incomplete transactions which must be undone, and redo-list consists of finished transactions that must be redone.  Recovery now continues as follows: 1. Scan log backwards from most recent record, stopping when <Ti start> records have been encountered for every Ti in undo-list.  During the scan, perform undo for each log record that belongs to a transaction in undo-list. 2. Locate the most recent <checkpoint L> record. 3. Scan log forwards from the <checkpoint L> record till the end of the log.  Database System Concepts, 5th Ed. During the scan, perform redo for each log record that belongs to a transaction on redo-list 17.31 ©Silberschatz, Korth and Sudarshan Example of Recovery  Go over the steps of the recovery algorithm on the following log: <T0 start> <T0, A, 0, 10> <T0 commit> <T1 start> <T1, B, 0, 10> <T2 start> /* Scan in Step 4 stops here */ <T2, C, 0, 10> <T2, C, 10, 20> <checkpoint {T1, T2}> <T3 start> <T3, A, 10, 20> <T3, D, 0, 10> <T3 commit> Database System Concepts, 5th Ed. 17.32 ©Silberschatz, Korth and Sudarshan Log Record Buffering  Log record buffering: log records are buffered in main memory, instead of of being output directly to stable storage.  Log records are output to stable storage when a block of log records in the buffer is full, or a log force operation is executed.  Log force is performed to commit a transaction by forcing all its log records (including the commit record) to stable storage.  Several log records can thus be output using a single output operation, reducing the I/O cost. Database System Concepts, 5th Ed. 17.33 ©Silberschatz, Korth and Sudarshan Log Record Buffering (Cont.)  The rules below must be followed if log records are buffered:  Log records are output to stable storage in the order in which they are created.  Transaction Ti enters the commit state only when the log record <Ti commit> has been output to stable storage.  Before a block of data in main memory is output to the database, all log records pertaining to data in that block must have been output to stable storage.  This rule is called the write-ahead logging or WAL rule – Strictly speaking WAL only requires undo information to be output Database System Concepts, 5th Ed. 17.34 ©Silberschatz, Korth and Sudarshan Database Buffering  Database maintains an in-memory buffer of data blocks  When a new block is needed, if buffer is full an existing block needs to be removed from buffer  If the block chosen for removal has been updated, it must be output to disk  As a result of the write-ahead logging rule, if a block with uncommitted updates is output to disk, log records with undo information for the updates are output to the log on stable storage first.  No updates should be in progress on a block when it is output to disk. Can be ensured as follows.  Before writing a data item, transaction acquires exclusive lock on block containing the data item  Lock can be released once the write is completed.   Such locks held for short duration are called latches. Before a block is output to disk, the system acquires an exclusive latch on the block  Ensures no update can be in progress on the block Database System Concepts, 5th Ed. 17.35 ©Silberschatz, Korth and Sudarshan Buffer Management (Cont.)   Database buffer can be implemented either  in an area of real main-memory reserved for the database, or  in virtual memory Implementing buffer in reserved main-memory has drawbacks:  Memory is partitioned before-hand between database buffer and applications, limiting flexibility.  Needs may change, and although operating system knows best how memory should be divided up at any time, it cannot change the partitioning of memory. Database System Concepts, 5th Ed. 17.36 ©Silberschatz, Korth and Sudarshan Buffer Management (Cont.)  Database buffers are generally implemented in virtual memory in spite of some drawbacks:  When operating system needs to evict a page that has been modified, to make space for another page, the page is written to swap space on disk.  When database decides to write buffer page to disk, buffer page may be in swap space, and may have to be read from swap space on disk and output to the database on disk, resulting in extra I/O!   Known as dual paging problem. Ideally when swapping out a database buffer page, operating system should pass control to database, which in turn outputs page to database instead of to swap space (making sure to output log records first)  Dual paging can thus be avoided, but common operating systems do not support such functionality. Database System Concepts, 5th Ed. 17.37 ©Silberschatz, Korth and Sudarshan Failure with Loss of Nonvolatile Storage  So far we assumed no loss of non-volatile storage  Technique similar to checkpointing used to deal with loss of non-volatile storage  Periodically dump the entire content of the database to stable storage  No transaction may be active during the dump procedure; a procedure similar to checkpointing must take place    Output all log records currently residing in main memory onto stable storage.  Output all buffer blocks onto the disk.  Copy the contents of the database to stable storage.  Output a record <dump> to log on stable storage. To recover from disk failure  restore database from most recent dump.  Consult the log and redo all transactions that committed after the dump Can be extended to allow transactions to be active during dump; known as fuzzy dump or online dump  Will study fuzzy checkpointing later Database System Concepts, 5th Ed. 17.38 ©Silberschatz, Korth and Sudarshan Advanced Recovery Algorithm Database System Concepts ©Silberschatz, Korth and Sudarshan See www.db-book.com for conditions on re-use Advanced Recovery Techniques  Support high-concurrency locking techniques, such as those used for B +-tree concurrency control  Operations like B+-tree insertions and deletions release locks early.   They cannot be undone by restoring old values (physical undo), since once a lock is released, other transactions may have updated the B +-tree.  Instead, insertions (resp. deletions) are undone by executing a deletion (resp. insertion) operation (known as logical undo). For such operations, undo log records should contain the undo operation to be executed   called logical undo logging, in contrast to physical undo logging. Redo information is logged physically (that is, new value for each write) even for such operations  Logical redo is very complicated since database state on disk may not be “operation consistent” Database System Concepts, 5th Ed. 17.40 ©Silberschatz, Korth and Sudarshan Advanced Recovery Techniques (Cont.)     Operation logging is done as follows: 1. When operation starts, log <Ti, Oj, operation-begin>. Here Oj is a unique identifier of the operation instance. 2. While operation is executing, normal log records with physical redo and physical undo information are logged. 3. When operation completes, <Ti, Oj, operation-end, U> is logged, where U contains information needed to perform a logical undo information. If crash/rollback occurs before operation completes:  the operation-end log record is not found, and  the physical undo information is used to undo operation. If crash/rollback occurs after the operation completes:  the operation-end log record is found, and in this case  logical undo is performed using U; the physical undo information for the operation is ignored. Redo of operation (after crash) still uses physical redo information. Database System Concepts, 5th Ed. 17.41 ©Silberschatz, Korth and Sudarshan Advanced Recovery Techniques (Cont.) Rollback of transaction Ti is done as follows:  Scan the log backwards 1. If a log record <Ti, X, V1, V2> is found, perform the undo and log a special redo-only log record <Ti, X, V1>. 2. If a <Ti, Oj, operation-end, U> record is found  Rollback the operation logically using the undo information U. – Updates performed during roll back are logged just like during normal operation execution. – At the end of the operation rollback, instead of logging an operationend record, generate a record <Ti, Oj, operation-abort>.  Database System Concepts, 5th Ed. Skip all preceding log records for Ti until the record <Ti, Oj operationbegin> is found 17.42 ©Silberschatz, Korth and Sudarshan Advanced Recovery Techniques (Cont.)  Scan the log backwards (cont.): 3. If a redo-only record is found ignore it 4. If a <Ti, Oj, operation-abort> record is found:  skip all preceding log records for Ti until the record <Ti, Oj, operation-begin> is found. 5. Stop the scan when the record <Ti, start> is found 6. Add a <Ti, abort> record to the log Some points to note:  Cases 3 and 4 above can occur only if the database crashes while a transaction is being rolled back.  Skipping of log records as in case 4 is important to prevent multiple rollback of the same operation. Database System Concepts, 5th Ed. 17.43 ©Silberschatz, Korth and Sudarshan Advanced Recovery Techniques(Cont,) The following actions are taken when recovering from system crash 1. Scan log forward from last < checkpoint L> record 1. Repeat history by physically redoing all updates of all transactions, 2. Create an undo-list during the scan as follows  undo-list is set to L initially  Whenever <Ti start> is found Ti is added to undo-list  Whenever <Ti commit> or <Ti abort> is found, Ti is deleted from undo-list This brings database to state as of crash, with committed as well as uncommitted transactions having been redone. Now undo-list contains transactions that are incomplete, that is, have neither committed nor been fully rolled back. Database System Concepts, 5th Ed. 17.44 ©Silberschatz, Korth and Sudarshan Advanced Recovery Techniques (Cont.) Recovery from system crash (cont.) 2. Scan log backwards, performing undo on log records of transactions found in undo-list.   Transactions are rolled back as described earlier.  When <Ti start> is found for a transaction Ti in undo-list, write a <Ti abort> log record.  Stop scan when <Ti start> records have been found for all Ti in undo-list This undoes the effects of incomplete transactions (those with neither commit nor abort log records). Recovery is now complete. Database System Concepts, 5th Ed. 17.45 ©Silberschatz, Korth and Sudarshan Advanced Recovery Techniques (Cont.)  Checkpointing is done as follows: 1. Output all log records in memory to stable storage 2. Output to disk all modified buffer blocks 3. Output to log on stable storage a < checkpoint L> record. Transactions are not allowed to perform any actions while checkpointing is in progress.  Fuzzy checkpointing allows transactions to progress while the most time consuming parts of checkpointing are in progress  Performed as described on next slide Database System Concepts, 5th Ed. 17.46 ©Silberschatz, Korth and Sudarshan Advanced Recovery Techniques (Cont.)  Fuzzy checkpointing is done as follows: 1. Temporarily stop all updates by transactions 2. Write a <checkpoint L> log record and force log to stable storage 3. Note list M of modified buffer blocks 4. Now permit transactions to proceed with their actions 5. Output to disk all modified buffer blocks in list M 6.   blocks should not be updated while being output  Follow WAL: all log records pertaining to a block must be output before the block is output Store a pointer to the checkpoint record in a fixed position last_checkpoint on disk When recovering using a fuzzy checkpoint, start scan from the checkpoint record pointed to by last_checkpoint  Log records before last_checkpoint have their updates reflected in database on disk, and need not be redone.  Incomplete checkpoints, where system had crashed while performing checkpoint, are handled safely Database System Concepts, 5th Ed. 17.47 ©Silberschatz, Korth and Sudarshan ARIES Recovery Algorithm Database System Concepts ©Silberschatz, Korth and Sudarshan See www.db-book.com for conditions on re-use ARIES   ARIES is a state of the art recovery method  Incorporates numerous optimizations to reduce overheads during normal processing and to speed up recovery  The “advanced recovery algorithm” we studied earlier is modeled after ARIES, but greatly simplified by removing optimizations Unlike the advanced recovery algorithm, ARIES 1. Uses log sequence number (LSN) to identify log records  Stores LSNs in pages to identify what updates have already been applied to a database page 2. Physiological redo 3. Dirty page table to avoid unnecessary redos during recovery 4. Fuzzy checkpointing that only records information about dirty pages, and does not require dirty pages to be written out at checkpoint time  Database System Concepts, 5th Ed. More coming up on each of the above … 17.49 ©Silberschatz, Korth and Sudarshan ARIES Optimizations  Physiological redo  Affected page is physically identified, action within page can be logical   Database System Concepts, 5th Ed. Used to reduce logging overheads – e.g. when a record is deleted and all other records have to be moved to fill hole » Physiological redo can log just the record deletion » Physical redo would require logging of old and new values for much of the page Requires page to be output to disk atomically – Easy to achieve with hardware RAID, also supported by some disk systems – Incomplete page output can be detected by checksum techniques, » But extra actions are required for recovery » Treated as a media failure 17.50 ©Silberschatz, Korth and Sudarshan ARIES Data Structures  Log sequence number (LSN) identifies each log record  Must be sequentially increasing  Typically an offset from beginning of log file to allow fast access   Easily extended to handle multiple log files Each page contains a PageLSN which is the LSN of the last log record whose effects are reflected on the page   To update a page:  X-latch the pag, and write the log record  Update the page  Record the LSN of the log record in PageLSN  Unlock page Page flush to disk S-latches page   Thus page state on disk is operation consistent – Required to support physiological redo PageLSN is used during recovery to prevent repeated redo  Thus ensuring idempotence Database System Concepts, 5th Ed. 17.51 ©Silberschatz, Korth and Sudarshan ARIES Data Structures (Cont.)  Each log record contains LSN of previous log record of the same transaction LSN TransId PrevLSN   LSN in log record may be implicit RedoInfo UndoInfo Special redo-only log record called compensation log record (CLR) used to log actions taken during recovery that never need to be undone  Also serve the role of operation-abort log records used in advanced recovery algorithm  Have a field UndoNextLSN to note next (earlier) record to be undone  Records in between would have already been undone  Required to avoid repeated undo of already undone actions LSN TransID UndoNextLSN RedoInfo Database System Concepts, 5th Ed. 17.52 ©Silberschatz, Korth and Sudarshan ARIES Data Structures (Cont.)   DirtyPageTable  List of pages in the buffer that have been updated  Contains, for each such page  PageLSN of the page  RecLSN is an LSN such that log records before this LSN have already been applied to the page version on disk – Set to current end of log when a page is inserted into dirty page table (just before being updated) – Recorded in checkpoints, helps to minimize redo work Checkpoint log record   Contains:  DirtyPageTable and list of active transactions  For each active transaction, LastLSN, the LSN of the last log record written by the transaction Fixed position on disk notes LSN of last completed checkpoint log record Database System Concepts, 5th Ed. 17.53 ©Silberschatz, Korth and Sudarshan ARIES Recovery Algorithm ARIES recovery involves three passes   Analysis pass: Determines  Which transactions to undo  Which pages were dirty (disk version not up to date) at time of crash  RedoLSN: LSN from which redo should start Redo pass:  Repeats history, redoing all actions from RedoLSN   RecLSN and PageLSNs are used to avoid redoing actions already reflected on page Undo pass:  Rolls back all incomplete transactions  Transactions whose abort was complete earlier are not undone – Key idea: no need to undo these transactions: earlier undo actions were logged, and are redone as required Database System Concepts, 5th Ed. 17.54 ©Silberschatz, Korth and Sudarshan ARIES Recovery: Analysis Analysis pass  Starts from last complete checkpoint log record  Reads in DirtyPageTable from log record  Sets RedoLSN = min of RecLSNs of all pages in DirtyPageTable   In case no pages are dirty, RedoLSN = checkpoint record’s LSN  Sets undo-list = list of transactions in checkpoint log record  Reads LSN of last log record for each transaction in undo-list from checkpoint log record Scans forward from checkpoint  .. On next page … Database System Concepts, 5th Ed. 17.55 ©Silberschatz, Korth and Sudarshan ARIES Recovery: Analysis (Cont.) Analysis pass (cont.)  Scans forward from checkpoint  If any log record found for transaction not in undo-list, adds transaction to undolist  Whenever an update log record is found   If transaction end log record found, delete transaction from undo-list  Keeps track of last log record for each transaction in undo-list   If page is not in DirtyPageTable, it is added with RecLSN set to LSN of the update log record May be needed for later undo At end of analysis pass:  RedoLSN determines where to start redo pass  RecLSN for each page in DirtyPageTable used to minimize redo work  All transactions in undo-list need to be rolled back Database System Concepts, 5th Ed. 17.56 ©Silberschatz, Korth and Sudarshan ARIES Redo Pass Redo Pass: Repeats history by replaying every action not already reflected in the page on disk, as follows:  Scans forward from RedoLSN. Whenever an update log record is found: 1. If the page is not in DirtyPageTable or the LSN of the log record is less than the RecLSN of the page in DirtyPageTable, then skip the log record 2. Otherwise fetch the page from disk. If the PageLSN of the page fetched from disk is less than the LSN of the log record, redo the log record NOTE: if either test is negative the effects of the log record have already appeared on the page. First test avoids even fetching the page from disk! Database System Concepts, 5th Ed. 17.57 ©Silberschatz, Korth and Sudarshan ARIES Undo Actions  When an undo is performed for an update log record  Generate a CLR containing the undo action performed (actions performed during undo are logged physicaly or physiologically).   Set UndoNextLSN of the CLR to the PrevLSN value of the update log record   Arrows indicate UndoNextLSN value ARIES supports partial rollback  Used e.g. to handle deadlocks by rolling back just enough to release reqd. locks  Figure indicates forward actions after partial rollbacks  1 CLR for record n noted as n’ in figure below 2 records 3 and 4 initially, later 5 and 6, then full rollback 3 Database System Concepts, 5th Ed. 4 4 ' 3 ' 5 17.58 6 6 ' 5 2 ' ' 1 ' ©Silberschatz, Korth and Sudarshan ARIES: Undo Pass Undo pass  Performs backward scan on log undoing all transaction in undo-list  Backward scan optimized by skipping unneeded log records as follows:  Next LSN to be undone for each transaction set to LSN of last log record for transaction found by analysis pass.  At each step pick largest of these LSNs to undo, skip back to it and undo it  After undoing a log record – For ordinary log records, set next LSN to be undone for transaction to PrevLSN noted in the log record – For compensation log records (CLRs) set next LSN to be undo to UndoNextLSN noted in the log record »  All intervening records are skipped since they would have been undo already Undos performed as described earlier Database System Concepts, 5th Ed. 17.59 ©Silberschatz, Korth and Sudarshan Other ARIES Features  Recovery Independence  Pages can be recovered independently of others   E.g. if some disk pages fail they can be recovered from a backup while other pages are being used Savepoints:  Transactions can record savepoints and roll back to a savepoint  Useful for complex transactions  Also used to rollback just enough to release locks on deadlock Database System Concepts, 5th Ed. 17.60 ©Silberschatz, Korth and Sudarshan Other ARIES Features (Cont.)  Fine-grained locking:  Index concurrency algorithms that permit tuple level locking on indices can be used   These require logical undo, rather than physical undo, as in advanced recovery algorithm Recovery optimizations: For example:  Dirty page table can be used to prefetch pages during redo  Out of order redo is possible:   redo can be postponed on a page being fetched from disk, and performed when page is fetched. Meanwhile other log records can continue to be processed Database System Concepts, 5th Ed. 17.61 ©Silberschatz, Korth and Sudarshan Remote Backup Systems Database System Concepts ©Silberschatz, Korth and Sudarshan See www.db-book.com for conditions on re-use Remote Backup Systems  Remote backup systems provide high availability by allowing transaction processing to continue even if the primary site is destroyed. Database System Concepts, 5th Ed. 17.63 ©Silberschatz, Korth and Sudarshan Remote Backup Systems (Cont.)  Detection of failure: Backup site must detect when primary site has failed   to distinguish primary site failure from link failure maintain several communication links between the primary and the remote backup. Transfer of control:  To take over control backup site first perform recovery using its copy of the database and all the long records it has received from the primary.  Thus, completed transactions are redone and incomplete transactions are rolled back.  When the backup site takes over processing it becomes the new primary  To transfer control back to old primary when it recovers, old primary must receive redo logs from the old backup and apply all updates locally. Database System Concepts, 5th Ed. 17.64 ©Silberschatz, Korth and Sudarshan Remote Backup Systems (Cont.)  Time to recover: To reduce delay in takeover, backup site periodically proceses the redo log records (in effect, performing recovery from previous database state), performs a checkpoint, and can then delete earlier parts of the log.  Hot-Spare configuration permits very fast takeover:  Backup continually processes redo log record as they arrive, applying the updates locally.  When failure of the primary is detected the backup rolls back incomplete transactions, and is ready to process new transactions.  Alternative to remote backup: distributed database with replicated data  Remote backup is faster and cheaper, but less tolerant to failure  more on this in Chapter 19 Database System Concepts, 5th Ed. 17.65 ©Silberschatz, Korth and Sudarshan Remote Backup Systems (Cont.)  Ensure durability of updates by delaying transaction commit until update is logged at backup; avoid this delay by permitting lower degrees of durability.  One-safe: commit as soon as transaction’s commit log record is written at primary   Two-very-safe: commit when transaction’s commit log record is written at primary and backup   Problem: updates may not arrive at backup before it takes over. Reduces availability since transactions cannot commit if either site fails. Two-safe: proceed as in two-very-safe if both primary and backup are active. If only the primary is active, the transaction commits as soon as is commit log record is written at the primary.  Better availability than two-very-safe; avoids problem of lost transactions in onesafe. Database System Concepts, 5th Ed. 17.66 ©Silberschatz, Korth and Sudarshan End of Chapter Database System Concepts ©Silberschatz, Korth and Sudarshan See www.db-book.com for conditions on re-use Block Storage Operations Database System Concepts, 5th Ed. 17.68 ©Silberschatz, Korth and Sudarshan Portion of the Database Log Corresponding to T0 and T1 Database System Concepts, 5th Ed. 17.69 ©Silberschatz, Korth and Sudarshan State of the Log and Database Corresponding to T0 and T1 Database System Concepts, 5th Ed. 17.70 ©Silberschatz, Korth and Sudarshan Portion of the System Log Corresponding to T0 and T1 Database System Concepts, 5th Ed. 17.71 ©Silberschatz, Korth and Sudarshan State of System Log and Database Corresponding to T0 and T1 Database System Concepts, 5th Ed. 17.72 ©Silberschatz, Korth and Sudarshan