{"id":113,"date":"2004-08-09T08:30:34","date_gmt":"2004-08-09T06:30:34","guid":{"rendered":""},"modified":"2007-09-06T07:39:05","modified_gmt":"2007-09-06T05:39:05","slug":"sun-solaris-svm-sds","status":"publish","type":"post","link":"http:\/\/www.lookit.org\/blog\/?p=113","title":{"rendered":"Sun Solaris SVM-SDS"},"content":{"rendered":"

Petit tutorial sur Sun Solaris Volume Manager, gestion RAID de Sun.
<\/p>\n

En cas de 2 disques : rajouter : set md:mirrored_root_flag=1 dans \/etc\/system , \u00e0 partir de 3 non( corum )<\/strong><\/em><\/u><\/p>\n

 <\/p>\n

Raid 0<\/u><\/em><\/strong> : concatenation<\/span> : on rempli A puis B<\/p>\n

Stripping : on partage data sur A et B<\/p>\n

 <\/p>\n

Ra<\/u><\/em><\/strong>id 1<\/span><\/u><\/em><\/strong>:mirroring, plexing<\/span><\/span><\/p>\n

<\/span><\/p>\n

Raid 0+1 :<\/span><\/span><\/span><\/u><\/em><\/strong> bof<\/span> :a_b_c<\/span>+a_b_c<\/span>stripping + mirroring<\/span><\/p>\n

0<\/span><\/span>0<\/span><\/p>\n

1<\/span><\/p>\n

 <\/p>\n

Raid 1+0 :<\/span><\/span><\/span><\/u><\/em><\/strong> bon : a_a<\/span>b_b<\/span>c_c<\/span><\/span><\/p>\n

11<\/span>1<\/span><\/span><\/p>\n

0<\/span><\/p>\n

<\/span><\/p>\n

Raid 2 :<\/span><\/span><\/u><\/em><\/strong>a + b + c = p_abc<\/span><\/span><\/p>\n

Raid 3 :<\/span><\/span><\/u><\/em><\/strong>a_d<\/span> + b_e<\/span> + c_f<\/span> = p_abc<\/span> + p_def<\/span><\/span><\/p>\n

Raid 4 : <\/u><\/em><\/strong>pareil que raid 3 : asynchrome<\/span><\/p>\n

Raid 5 :<\/u><\/em><\/strong>pareil que raid 3_4 : gestion des priorités : partage<\/p>\n

Raid 6 : <\/u><\/em><\/strong>Double parité<\/p>\n

 <\/p>\n

Raid soft<\/u><\/em><\/strong> :solaris<\/span> + svm<\/span> + VM<\/p>\n

V<\/p>\n

Bus scsi<\/span><\/p>\n

V<\/p>\n

( Disks<\/span><\/span> ) ou ( disks<\/span> + cache)<\/p>\n

 <\/p>\n

Raid Hard<\/u><\/em><\/strong> : solaris<\/span> ( c1t3<\/span>… )<\/p>\n

V<\/p>\n

Bus scsi<\/span> 1 ( ou<\/span> plusieurs Bus scsi<\/span> en miroir )<\/p>\n

VV<\/p>\n

Controleur<\/span> raid > avec disks<\/span> ( appelés<\/span> LUN ou DEVICE )<\/p>\n

 <\/p>\n

Raid Soft + Hard :<\/span><\/span><\/span><\/u><\/em><\/strong> plaiding<\/span> ( miroir<\/span> ) entre<\/span> 2 controleur<\/span> raid.<\/span><\/p>\n

<\/span><\/p>\n

SVM: <\/u><\/em><\/strong><\/p>\n

– 3 databases<\/span> minimum.<\/p>\n

– Nécessaire pour booter au minimum : la moitié + 1 db<\/span> ( notre<\/span> paramétrable ds<\/span> \/etc<\/span>\/system )<\/p>\n

– Si la moitié tombe : c’est ok … Si < à la moitié : PROBLEME<\/p>\n

– Jusqu'à Sol8 inclus : infos des metadbs<\/span> ds<\/span> \/etc<\/span>\/system<\/p>\n

– Les devices<\/span> sont créés ds<\/span> md.cf \/dev<\/span>\/md\/rdsk<\/span>* et \/dev<\/span>\/md\/dsk<\/span>*<\/p>\n

– BACKUP a<\/span> faire avant crash : \/etc<\/span>\/vfstab<\/span>, \/etc<\/span>\/system, backup des prtvtoc<\/span>, backup de la conf<\/span> SVM<\/p>\n

– NE PAS TOUCHER AUX FICHIERS : \/etc<\/span>\/lvm<\/span>\/mddb.cf, \/etc<\/span>\/lvm<\/span>\/md.cf, \/kernel<\/span>\/drv<\/span>\/md.conf<\/span> ( config<\/span> de SUN ).<\/p>\n

– ON PEUT TOUCHER : \/etc<\/span>\/lvm<\/span>\/md.tab.<\/p>\n

– metaparam<\/span> : paramétrer : l’ordre de synchro des disques, méthode d’écriture\/lecture sur les disques : metapram<\/span> -5 d20 : synchro en 5° place … etc.<\/p>\n

 <\/p>\n

Concat<\/span><\/u><\/em><\/strong><\/span><\/span> ,<\/span><\/u><\/em><\/strong><\/span> Stripping : <\/span><\/u><\/em><\/strong><\/p>\n

Metainit<\/span><\/span> d11 1 3 c3t1d0s5 c3t2d0s5 c3t3d0s5 –i<\/span> 256k <\/span><\/p>\n

>1 : 1 pour stripe<\/span>, 2ou + : concat<\/span><\/p>\n

>3 : partition<\/span> du concat<\/span> ou stripe<\/span><\/p>\n

>-i : taille des blocs écris<\/p>\n

 <\/p>\n

Concat<\/span> + stripping : <\/p>\n

P1 P2 P3<\/p>\n

P3 P4 P5<\/p>\n

Metainit<\/span><\/span> d132 2 3 p1 p2 p3 –i<\/span> 256k<\/strong> p4 p5 p6 –i<\/span> 256k<\/strong><\/span><\/p>\n

<\/span><\/p>\n

Mirroring :<\/span><\/u><\/em><\/strong><\/span><\/u><\/em><\/strong><\/p>\n

<\/span><\/p>\n

Crée<\/span><\/em><\/strong><\/span> chq volume :<\/span> <\/span><\/em><\/strong><\/p>\n

Metainit<\/span><\/span> d11 1 1<\/span> c3t1d0s0<\/span><\/p>\n

Metainit<\/span><\/span> d12 1 1<\/span> c3t2d0s0<\/span><\/p>\n

Créer<\/span><\/em><\/strong><\/span> master volume :<\/span><\/span><\/em><\/strong><\/p>\n

Metainit<\/span><\/span> d10 md<\/span> d11<\/span><\/p>\n

Metattach<\/span><\/span> d10 12<\/span><\/p>\n

Metastat<\/span><\/span><\/p>\n

<\/span><\/p>\n

DISK EN PANNE :<\/u><\/em><\/strong><\/p>\n

Metadb<\/span> –i : si majuscules : problème dans la base de données.<\/p>\n

>> Jusqu’à sol8 : <\/strong><\/p>\n

-détache le disque<\/p>\n

-change le disque<\/p>\n

-format puis création des partitions ( fmthard<\/span><\/span> )<\/p>\n

–metareplace<\/span><\/span> –e d10 p1<\/span><\/p>\n

–metareplace<\/span><\/span> –e d20 p2<\/span><\/p>\n

–metareplace<\/span> pour chq<\/span> partition incluse.<\/p>\n

–metadb<\/span><\/span> –d oldpartdbs7<\/strong><\/span><\/p>\n

–metadb<\/span><\/span> –ac 2 newpartdbs7<\/strong><\/span><\/p>\n

>>Sol10 :<\/strong><\/p>\n

-détache le disque<\/p>\n

-change le disque<\/p>\n

-format puis création des partitions ( fmthard<\/span><\/span> )<\/p>\n

–metadevadm<\/span> –u NEWDISK<\/strong><\/p>\n

–metareplace<\/span><\/span> –e d10 p1<\/span><\/p>\n

–metareplace<\/span><\/span> –e d20 p2<\/span><\/p>\n

–metareplace<\/span> pour chq<\/span> partition incluse.<\/p>\n

–metadb<\/span><\/span> –d oldpartdbs7<\/strong><\/span><\/p>\n

–metadb<\/span><\/span> –ac 2 newpartdbs7<\/strong><\/span><\/p>\n

<\/span><\/p>\n

OFFLINE et ONLINE : activer ou désactiver des devices<\/span><\/u><\/em><\/strong><\/p>\n

Metaoffline<\/span> d10 d12<\/p>\n

Metaonline<\/span> d10 12<\/p>\n

Metastat<\/span><\/p>\n

Utilisation : faire un backup par exemple, snapshot<\/span> …<\/p>\n

 <\/p>\n

HOT SPARE POOL : disque de recharge en cas de panne.<\/u><\/em><\/strong><\/p>\n

Exemple :<\/p>\n

1)<\/p>\n

d50<\/span> avec d51 et d52 : 2 x 1Go<\/p>\n

metainit<\/span><\/span><\/span> d51 1 1<\/span> c3t1d0s0<\/span><\/p>\n

metainit<\/span><\/span><\/span> d52 1 1<\/span> c3t1d0s0<\/span><\/p>\n

metainit<\/span><\/span><\/span> d50 –m d51<\/span><\/p>\n

metattach<\/span><\/span><\/span> d50 d52<\/span><\/p>\n

newfs<\/span><\/span><\/span> \/dev\/md\/rdsk\/d50<\/span><\/p>\n

mount<\/span><\/span> \/dev\/md\/dsk\/d50 \/miroir<\/span><\/span><\/p>\n

2)<\/span><\/p>\n

hsp000 :<\/span><\/span> 2 x 1Go : c3t3d0s6 \/\/ c3t4d0s6<\/span><\/p>\n

metainit<\/span><\/span><\/span> hsp000 c3t3d0s6<\/span><\/p>\n

metahs<\/span><\/span><\/span> –a hsp000 c3t4d0s6 ( ajout<\/span> d’un hot spare )<\/span><\/p>\n

metahs<\/span><\/span><\/span> –i<\/span> : info globale<\/span><\/span><\/p>\n

metaparam<\/span><\/span> –h hsp000 d51 <<< lier hsp<\/span> aux devices<\/span><\/p>\n

metaparam<\/span><\/span> –h hsp000 d52 <<< lier hsp<\/span> aux devices<\/span><\/p>\n

3) <\/p>\n

le<\/span> disque casse<\/p>\n

4)<\/p>\n

Changement disque, format, fmthard<\/span><\/span> ,<\/span> partition …<\/p>\n

Metadevadm<\/span><\/span> –u c3t1d0s0 < old disk<\/span><\/p>\n

Metareplace<\/span> –e d50 c3t1d0s0 < nvl<\/span> partition<\/p>\n

Resync<\/span> et puis c’est OK.<\/p>\n

5)<\/p>\n

Enable<\/span> nvl<\/span> partitions :<\/p>\n

Metahs<\/span> –e c3t3d0s0<\/p>\n

Metahs<\/span> –e chq<\/span> partion<\/span> <\/p>\n

6)<\/span><\/p>\n

Metadb<\/span> –d c3t1d0s7 < ancienne partition de db<\/span><\/p>\n

Metadb<\/span> –ac<\/span> 2 c3t1d0s7 < nvl<\/span> partition de db<\/span><\/p>\n

 <\/p>\n

RAID 5 :<\/span><\/span><\/u><\/em><\/strong><\/p>\n

Metainit<\/span><\/span> d30 –r c3t1d0s0 c3t2d0s0 c3t3d0s0<\/span><\/p>\n

<\/span><\/p>\n

SOFT PARTITIONS: <\/u><\/em><\/strong><\/p>\n

http:\/\/sysunconfig.net\/unixtips\/soft-partitions.html<\/span><\/a><\/span><\/p>\n

créer des partitions virtuelles ( plus<\/span> que 7 partitions )<\/p>\n

 <\/p>\n

REPARER DB SVM :<\/u><\/em><\/strong><\/p>\n

Si on a 3 partitions s7 dont 2 HS : crash du serveur<\/p>\n

>> reboot « boot –r »<\/span><\/p>\n

>> metadb<\/span><\/span> –d part_hs1<\/span><\/p>\n

>> metadb<\/span><\/span> –d part_hs2<\/span><\/p>\n

Voir ds<\/span> \/etc<\/span>\/vfstab<\/span> pour enlever le dXX<\/span><\/p>\n

>>Sol10 : svcs<\/span> –x : service qui plante après lecture du \/etc<\/span>\/vfstab<\/span><\/p>\n

 <\/p>\n

MIRRORER DISQUE SYSTEME :<\/span><\/span><\/u><\/em><\/strong><\/p>\n

Metainit<\/span><\/span> –f d1 1 1<\/span> c0t0d0s0<\/span><\/p>\n

Metainit<\/span><\/span> –f d11 1 1<\/span> c0t0d0s1<\/span><\/p>\n

Metainit<\/span><\/span> –f d61 1 1<\/span> c0t0d0s6<\/span><\/p>\n

Metainit<\/span><\/span> –f d0 –m d1<\/span><\/p>\n

Metainit<\/span><\/span> –f d10 –m d11<\/span><\/p>\n

Metainit<\/span> –f d60 –m d61<\/p>\n

 <\/p>\n

Cp \/etc<\/span>\/vfstab<\/span> \/etc<\/span>\/vfstab.ori<\/p>\n

Cp \/etc<\/span>\/system \/etc<\/span>\/system.ori<\/p>\n

Editer \/etc<\/span>\/vfstab<\/span><\/p>\n

Metaroot<\/span> d0 puis reboot<\/p>\n

 <\/p>\n

DISK SET : pour clusters<\/u><\/em><\/strong><\/p>\n

 <\/p>\n

S1S2<\/p>\n

C1 C2C1 C2<\/p>\n

 <\/p>\n

C1 = o et x<\/p>\n

C2 = o o<\/span> et x x<\/span><\/p>\n

>>> 2 groupes de disks<\/span> : A et B appelés : « diskset<\/span> »<\/p>\n

A : o o<\/span> o<\/span> o<\/span> ( 4<\/span> disques )<\/p>\n

B : x x<\/span> x<\/span> x<\/span>( 4<\/span> disques )<\/p>\n

Chaques<\/span> disques : avec svm<\/span> doit avoir une 1° partition N°7 sur le 1° cylindre et rien sur le reste du disque ( pas<\/span> de partition 2 ).<\/p>\n

 <\/p>\n

C1 et C2 : il faut que ce soit les mêmes noms de liens ( c0t0<\/span>… ) sur les 2 serveurs<\/strong><\/p>\n

EN cluster SUN : il n’y a plus d’obligation de mêmes noms : il y a les disques virtuels : « did<\/span> » : device<\/span> id.<\/p>\n

 <\/p>\n

>>> Metaset<\/span> –s A<\/span><\/p>\n

>>> Metaset<\/span> –s B<\/span><\/p>\n

TOUJOURS metre<\/span> “-s” pour gérer les disksets<\/span> : \/dev<\/span>\/md\/A\/rdsk<\/span>\/d0<\/p>\n

 <\/p>\n

Pour ajouter des diskset<\/span> sur la 2° machine : il faut que le service en sol10 : « metainit<\/span> » soit chargé. En solaris<\/span> 8 : rien à faire.<\/p>\n

 <\/p>\n

Exemple : <\/strong><\/p>\n

SUR S1 :<\/strong><\/p>\n

#Metaset<\/span> : info sur les disksets<\/span>.<\/p>\n

#metaset<\/span> –s A –a –h s1 ( s1<\/span> étant le 1° serveur où l’on déclare le diskset<\/span> )<\/p>\n

#metaset<\/span> –s A –a –h s2 :<\/span> d° serveur<\/span><\/span><\/p>\n

#metaset<\/span> –s A –a c3t1d0<\/span><\/p>\n

#metaset<\/span> –s A –a c3t2d0<\/span><\/p>\n

#metaset<\/span> –s A :<\/span> on voit<\/span> le résultat<\/span><\/span><\/p>\n

#metaset<\/span> –s A –r ( “<\/span>-r” : release du diskset<\/span> A pour le 2° serveur )<\/p>\n

SUR S2 :<\/span> <\/span><\/strong><\/p>\n

<\/p>\n

#metaset<\/span> –s A<\/span> -t<\/span><\/p>\n

<\/p>\n

ANNOTATION :<\/p>\n

<\/p>\n

comment<\/p>\n

faire du sds sur 2 disques<\/strong> et garantir le boot en cas de panne<\/strong> d'un des disques,<\/p>\n

voici une option a placer dans le \/etc\/system pour contourner l'algorithme de <\/p>\n

consensus de majorité des metadb :<\/p>\n

set md:mirrored_root_flag=1<\/p>\n

Avec ça le systeme boot normalement, meme avec un seul disque systeme sous sds. <\/p>\n","protected":false},"excerpt":{"rendered":"

Petit tutorial sur Sun Solaris Volume Manager, gestion RAID de Sun.<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"jetpack_post_was_ever_published":false,"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":false,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2}},"categories":[8],"tags":[],"class_list":["post-113","post","type-post","status-publish","format-standard","hentry","category-informatique"],"jetpack_publicize_connections":[],"jetpack_featured_media_url":"","jetpack_shortlink":"https:\/\/wp.me\/pWrTo-1P","jetpack_sharing_enabled":true,"jetpack_likes_enabled":true,"_links":{"self":[{"href":"http:\/\/www.lookit.org\/blog\/index.php?rest_route=\/wp\/v2\/posts\/113","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.lookit.org\/blog\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.lookit.org\/blog\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.lookit.org\/blog\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.lookit.org\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=113"}],"version-history":[{"count":0,"href":"http:\/\/www.lookit.org\/blog\/index.php?rest_route=\/wp\/v2\/posts\/113\/revisions"}],"wp:attachment":[{"href":"http:\/\/www.lookit.org\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=113"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.lookit.org\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=113"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.lookit.org\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=113"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}