[Ubuntu] Mdadm et Raid 10 logiciel - erreur au boot

Mdadm et Raid 10 logiciel - erreur au boot [Ubuntu] - Installation - Linux et OS Alternatifs

Marsh Posté le 24-04-2006 à 16:13:47    

Bonjour messieurs,
 
J'ai récemment configuré un Raid 10 logiciel sur une ubuntu, avec 6 disques de 300Go. Donc deux raid 0 de 600, md0 et md1, le tout en Raid1 en tant que md2.
 
Quand je fais un cat /proc/mdstat, tout va bien :
 


root@*:~# cat /proc/mdstat
Personalities : [raid0] [raid1]
md2 : active raid1 md0[0] md1[1]
      578275520 blocks [2/2] [UU]
 
md1 : active raid0 sdc3[0] sdd3[1]
      578275584 blocks 64k chunks
 
md0 : active raid0 sda3[0] sdb3[1]
      578275584 blocks 64k chunks
 
unused devices: <none>


 
Le monitor de mdadm me dit bien que mon raid1 va super bien.
 
Cependant, au boot, j'ai bien un message qui me dit "Starting Raid Devices : Failed". Quand j'active bootlogd ( yay ), je retrouve bien l'erreur sur la première ligne ( si quelqu'un sait comment niquer ces codes de couleur, je lui roule un palot).


Mon Apr 24 14:59:40 2006: ^[[74G[^[[31mfail^[[39;49m]
Mon Apr 24 14:59:40 2006: ^[[74G[ ok ]ng disc parameters...       ^[[80G
Mon Apr 24 14:59:40 2006:  * Checking root file system...       ^[[80G /: clean,
74841/366528 files, 420964/732957 blocks
Mon Apr 24 14:59:40 2006: ^[[74G[ ok ]
Mon Apr 24 14:59:40 2006: ^[[74G[ ok ]ing modules...       ^[[80G
Mon Apr 24 14:59:41 2006: ^[[74G[ ok ]up ifupdown...       ^[[80G
Mon Apr 24 14:59:41 2006: ^[[74G[ ok ]ng module dependencies...       ^[[80G
Mon Apr 24 14:59:43 2006: ^[[74G[ ok ]odules...       ^[[80G
Mon Apr 24 14:59:44 2006: ^[[74G[ ok ]he system clock...       ^[[80G
Mon Apr 24 14:59:46 2006: ^[[74G[ ok ]p LVM Volume Groups...       ^[[80G
Mon Apr 24 14:59:46 2006: ^[[74G[ ok ]Enterprise Volume Management System...    
  ^[[80G
Mon Apr 24 14:59:47 2006: ^[[74G[ ok ]all file systems...       ^[[80G
Mon Apr 24 14:59:47 2006:  * Mounting local filesystems...       ^[[80G mount:
you didn't specify a filesystem type for /dev/md2
Mon Apr 24 14:59:47 2006:        I will try type ext3
Mon Apr 24 14:59:47 2006: /dev/md2 on /raid type ext3 (rw)
Mon Apr 24 14:59:47 2006: ^[[74G[ ok ]
Mon Apr 24 14:59:48 2006: ^[[74G[ ok ] resolver state...       ^[[80G
Mon Apr 24 14:59:48 2006: ^[[74G[ ok ]p networking...       ^[[80G
Mon Apr 24 14:59:48 2006: ^[[74G[ ok ]esktop files...       ^[[80G
Mon Apr 24 14:59:49 2006: ^[[74G[ ok ]hotplug subsystem...       ^[[80G
Mon Apr 24 14:59:53 2006: ^[[74G[ ok ]ng network interfaces...       ^[[80G
Mon Apr 24 15:00:02 2006: ^[[74G[ ok ]or network interface to come up...      
^[[80G
Mon Apr 24 15:00:02 2006: ^[[74G[ ok ]p ALSA...       ^[[80G
Mon Apr 24 15:00:02 2006: ^[[74G[ ok ]he system clock...       ^[[80G
Mon Apr 24 15:00:03 2006: ^[[74G[ ok ]zing clock to ntp.ubuntulinux.org...      
^[[80G
Mon Apr 24 15:00:03 2006: ^[[74G[ ok ]ing random number generator...      
^[[80G
Mon Apr 24 15:00:03 2006: ^[[74G[ ok ]p X server socket directory...      
^[[80G
Mon Apr 24 15:00:03 2006: ^[[74G[ ok ]p ICE socket directory...       ^[[80G
Mon Apr 24 15:00:03 2006:  * Entering runlevel: 2
Mon Apr 24 15:00:04 2006: ^[[74G[ ok ]CPI modules...       ^[[80G
Mon Apr 24 15:00:04 2006: ^[[74G[ ok ]ACPI services...       ^[[80G
Mon Apr 24 15:00:04 2006: ^[[74G[ ok ]system log daemon...       ^[[80G
Mon Apr 24 15:00:05 2006: ^[[74G[ ok ]kernel log daemon...       ^[[80G
Mon Apr 24 15:00:05 2006: ^[[74G[ ok ]system message bus...       ^[[80G
Mon Apr 24 15:00:05 2006: ^[[74G[ ok ]Hardware abstraction layer:        ^[[80G
Mon Apr 24 15:00:06 2006: ^[[74G[ ok ]ng up general console font...       ^[[80G
Mon Apr 24 15:00:07 2006: ^[[74G[ ok ]6n^[D^[[1F^[[1X^[%G^[[9;30]^[[14;30] *
Starting GNOME Display Manager...       ^[[80G
Mon Apr 24 15:00:07 2006: ^[[74G[ ok ]Common Unix Printing System: cupsd      
^[[80G
Mon Apr 24 15:00:08 2006: ^[[74G[ ok ]HP Linux Printing and Imaging System      
^[[80G
Mon Apr 24 15:00:11 2006:  * PCMCIA not present
Mon Apr 24 15:00:11 2006:  * Starting powernowd...        ^[[80G  * CPU
frequency scaling not supported
Mon Apr 24 15:00:11 2006: ^[[74G[ ok ]
Mon Apr 24 15:00:12 2006: ^[[74G[ ok ]Samba daemons..       ^[[80G
Mon Apr 24 15:00:12 2006: ^[[74G[ ok ]OpenBSD Secure Shell server...      
^[[80G
Mon Apr 24 15:00:12 2006: ^[[74G[ ok ]Bluetooth services...       ^[[80G  hcid
sdpd
Mon Apr 24 15:00:13 2006: ^[[74G[ ok ]RAID monitoring services...       ^[[80G
Mon Apr 24 15:00:13 2006: ^[[74G[ ok ]anac(h)ronistic cron: anacron       ^[[80G
Mon Apr 24 15:00:13 2006: ^[[74G[ ok ]deferred execution scheduler...      
^[[80G
Mon Apr 24 15:00:13 2006: ^[[74G[ ok ]periodic command scheduler...       ^[[80G
Mon Apr 24 15:00:13 2006: ^[[74G[ ok ]battery state...       ^[[80G


 
Et je ne vois strictement rien d'anormal dans le dmesg


[4294679.373000] md: md driver 0.90.1 MAX_MD_DEVS=256, MD_SB_DISKS=27
[4294679.377000] md: raid0 personality registered as nr 2
[4294679.691000] devfs_mk_dev: could not append to parent for md/0
[4294679.694000] md: md0 stopped.
[4294679.695000] md: bind<sdb3>
[4294679.695000] md: bind<sda3>
[4294679.695000] md0: setting max_sectors to 128, segment boundary to 32767
[4294679.695000] raid0: looking at sda3
[4294679.695000] raid0:   comparing sda3(289137792) with sda3(289137792)
[4294679.695000] raid0:   END
[4294679.695000] raid0:   ==> UNIQUE
[4294679.695000] raid0: 1 zones
[4294679.695000] raid0: looking at sdb3
[4294679.695000] raid0:   comparing sdb3(289137792) with sda3(289137792)
[4294679.695000] raid0:   EQUAL
[4294679.695000] raid0: FINAL 1 zones
[4294679.695000] raid0: done.
[4294679.695000] raid0 : md_size is 578275584 blocks.
[4294679.695000] raid0 : conf->hash_spacing is 578275584 blocks.
[4294679.695000] raid0 : nb_zone is 1.
[4294679.695000] raid0 : Allocating 4 bytes for hash.
[4294679.709000] devfs_mk_dev: could not append to parent for md/1
[4294679.712000] md: md1 stopped.
[4294679.713000] md: bind<sdd3>
[4294679.713000] md: bind<sdc3>
[4294679.713000] md1: setting max_sectors to 128, segment boundary to 32767
[4294679.713000] raid0: looking at sdc3
[4294679.713000] raid0:   comparing sdc3(289137792) with sdc3(289137792)
[4294679.713000] raid0:   END
[4294679.713000] raid0:   ==> UNIQUE
[4294679.713000] raid0: 1 zones
[4294679.713000] raid0: looking at sdd3
[4294679.713000] raid0:   comparing sdd3(289137792) with sdc3(289137792)
[4294679.713000] raid0:   EQUAL
[4294679.713000] raid0: FINAL 1 zones
[4294679.713000] raid0: done.
[4294679.713000] raid0 : md_size is 578275584 blocks.
[4294679.713000] raid0 : conf->hash_spacing is 578275584 blocks.
[4294679.713000] raid0 : nb_zone is 1.
[4294679.713000] raid0 : Allocating 4 bytes for hash.
[4294679.747000] Attempting manual resume
[4294679.748000] swsusp: Suspend partition has wrong signature?
[4294679.764000] kjournald starting.  Commit interval 5 seconds
[4294679.764000] EXT3-fs: mounted filesystem with ordered data mode.
[4294681.321000] devfs_mk_dev: could not append to parent for md/2
[4294681.328000] md: md2 stopped.
[4294681.337000] md: bind<md1>
[4294681.337000] md: bind<md0>
[4294681.632000] md: raid1 personality registered as nr 3
[4294681.632000] raid1: raid set md2 active with 2 out of 2 mirrors


 
Une idée de comment détecter / débugguer le souci ?
 
Merci d'avance :hello:


Message édité par Tetedeiench le 24-04-2006 à 16:15:04
Reply

Marsh Posté le 24-04-2006 à 16:13:47   

Reply

Marsh Posté le 24-04-2006 à 17:31:25    

En fait, si y avait un moyen de savoir pourquoi je l'ai en [failed] quelque part ( pas trouvé de log qui veuille dire quelque chose...), ce serait nickel.
 
C'est ptêt un problème de monitoring - y a un moyen de vérifier ?

Reply

Marsh Posté le 24-04-2006 à 21:51:35    

Réglé ! En fait, la simple présence d'un fichier de conf dans /etc/mdadm/mdadm.conf implique le lancement, au boot de la commande mdadm -A -s qui, si elle ne trouve pas de raid défini dans le fichier de conf, renvoie -1...
 
Réglé en virant ce fichier qui ne me sert à rien, finalement.

Reply

Marsh Posté le 17-05-2006 à 17:44:58    

:hello:  
 
Tu sais faire un hdparm -tT /dev/md2 juste par curiosite ?
 
Avec 4 disques, tu devrais vraiment essayer le raid5. Chez moi c'est hyper performant alors que j'ai que trois disques (et qu'ils sont de modele, capacite et taille differente) [:spamafote]

Reply

Sujets relatifs:

Leave a Replay

Make sure you enter the(*)required information where indicate.HTML code is not allowed