From mboxrd@z Thu Jan 1 00:00:00 1970 From: Roy Sigurd Karlsbakk Subject: Re: Expanding raid0 Date: Fri, 21 Dec 2012 18:32:53 +0100 (CET) Message-ID: <12155358.8.1356111173286.JavaMail.root@zimbra> References: Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: Sender: linux-raid-owner@vger.kernel.org To: Kubes Cc: linux-raid List-Id: linux-raid.ids well, I guess it should work, in theory. I just tried with loopback dev= ices (1GB each). This is on Ubuntu Precise 12.04.1, kernel 3.2.0-35-gen= eric. Got an OOPS, which triggered a panic (kernel.panic_on_oops =3D 1)= which triggered a reboot (kernel.panic =3D 1). It might work, but not = here - other RAID stuff works well on this box See below for logs roy root@smilla:/raid/tmp/testraid# mdadm --create /dev/md1 --level=3D0 --c= hunk=3D256 --raid-devices=3D2 /dev/loop0 /dev/loop1 mdadm: Defaulting to version 1.2 metadata mdadm: array /dev/md1 started. root@smilla:/raid/tmp/testraid# cat /proc/mdstat=20 Personalities : [raid6] [raid5] [raid4] [linear] [multipath] [raid0] [r= aid1] [raid10]=20 md1 : active raid0 loop1[1] loop0[0] 2096640 blocks super 1.2 256k chunks =20 root@smilla:/raid/tmp/testraid# mkfs.xfs /dev/md1 =2E.. (copy 200MB into the filesystem) root@smilla:/mnt/raidtest# mdadm -G /dev/md1 --raid-disks=3D3 --add /de= v/loop2 mdadm: level of /dev/md1 changed to raid4 mdadm: added /dev/loop2 mdadm: Need to backup 1536K of critical section.. mdadm: Cannot start reshape for /dev/md1 mdadm: aborting level change root@smilla:/mnt/raidtest# Timeout, server 192.168.10.254 not respondin= g. seems like an OOPS was triggered Dec 21 18:25:53 smilla kernel: [ 4742.265318] md/raid:md1: device loop1= operational as raid disk 1 Dec 21 18:25:53 smilla kernel: [ 4742.265321] md/raid:md1: device loop0= operational as raid disk 0 Dec 21 18:25:53 smilla kernel: [ 4742.265587] md/raid:md1: allocated 32= 28kB Dec 21 18:25:53 smilla kernel: [ 4742.284092] md/raid:md1: raid level 4= active with 2 out of 3 devices, algorithm 5 Dec 21 18:25:53 smilla kernel: [ 4742.285250] RAID conf printout: Dec 21 18:25:53 smilla kernel: [ 4742.285252] --- level:4 rd:3 wd:2 Dec 21 18:25:53 smilla kernel: [ 4742.285254] disk 0, o:1, dev:loop0 Dec 21 18:25:53 smilla kernel: [ 4742.285255] disk 1, o:1, dev:loop1 Dec 21 18:25:54 smilla kernel: [ 4742.987274] md: bind Dec 21 18:25:55 smilla kernel: [ 4743.426366] md/raid0:md1: bad disk nu= mber -1 - aborting! Dec 21 18:25:55 smilla kernel: [ 4743.444444] md/raid0:md1: bad disk nu= mber -1 - aborting! Dec 21 18:25:56 smilla kernel: [ 4743.451231] BUG: unable to handle ker= nel NULL pointer dereference at 0000000000000010 Dec 21 18:25:56 smilla kernel: [ 4743.452474] IP: [] = raid0_make_request+0x69/0x1c0 [raid0] Dec 21 18:25:56 smilla kernel: [ 4743.453608] PGD 137668067 PUD 1314240= 67 PMD 0=20 Dec 21 18:25:56 smilla kernel: [ 4743.454283] Oops: 0000 [#1] SMP=20 Dec 21 18:25:56 smilla kernel: [ 4743.454283] CPU 0=20 Dec 21 18:25:56 smilla kernel: [ 4743.454283] Modules linked in: xfs eb= table_nat ebtables ipt_MASQUERADE iptable_nat xt_CHECKSUM iptable_mangl= e kvm_amd kvm dm_crypt xt_hl ip6t_rt ib_iser rdma_cm nf_conntrack_ipv6 = nf_defrag_ipv6 ib_cm iw_cm ib_sa ib_mad ib_core ipt_REJECT ib_addr iscs= i_tcp libiscsi_tcp libiscsi scsi_transport_iscsi xt_limit xt_tcpudp xt_= addrtype xt_state ip6table_filter ip6_tables nf_conntrack_netbios_ns nf= _conntrack_broadcast nf_nat_ftp nf_nat nf_conntrack_ipv4 nf_defrag_ipv4= nf_conntrack_ftp nf_conntrack iptable_filter ip_tables x_tables snd_hd= a_codec_via snd_hda_intel snd_hda_codec snd_hwdep snd_pcm bridge stp sn= d_seq_midi snd_rawmidi snd_seq_midi_event arc4 snd_seq rt61pci crc_itu_= t rt2x00pci rt2x00lib mac80211 snd_timer snd_seq_device cfg80211 eeprom= _93cx6 sp5100_tco snd soundcore psmouse snd_page_alloc uvcvideo i2c_pii= x4 videodev v4l2_compat_ioctl32 serio_raw edac_core k10temp edac_mce_am= d ext2 mac_hid w83627ehf hwmon_vid lp parport raid10 raid1 raid0 multip= ath linear btrfs zlib_deflate lib Dec 21 18:25:56 smilla kernel: crc32c raid456 async_pq async_xor xor as= ync_memcpy async_raid6_recov raid6_pq async_tx radeon ttm drm_kms_helpe= r drm i2c_algo_bit sata_sil24 pata_atiixp usbhid hid r8169 Dec 21 18:25:56 smilla kernel: [ 4743.454283]=20 Dec 21 18:25:56 smilla kernel: [ 4743.454283] Pid: 16410, comm: blkid N= ot tainted 3.2.0-35-generic #55-Ubuntu To Be Filled By O.E.M. To Be Fil= led By O.E.M./M3A UCC Dec 21 18:25:56 smilla kernel: [ 4743.454283] RIP: 0010:[] [] raid0_make_request+0x69/0x1c0 [raid0] Dec 21 18:25:56 smilla kernel: [ 4743.454283] RSP: 0018:ffff8800be3918c= 8 EFLAGS: 00010202 Dec 21 18:25:56 smilla kernel: [ 4743.454283] RAX: 0000000000000001 RBX= : ffff88002bcabc00 RCX: 0000000000000190 Dec 21 18:25:56 smilla kernel: [ 4743.454283] RDX: 0000000000000001 RSI= : 0000000000000200 RDI: 0000000000000000 Dec 21 18:25:56 smilla kernel: [ 4743.454283] RBP: ffff8800be3918e8 R08= : 0000000000000200 R09: 0000000000000100 Dec 21 18:25:56 smilla kernel: [ 4743.454283] R10: ffff8801877acf08 R11= : 0000000000000000 R12: ffff88021f0566c0 Dec 21 18:25:56 smilla kernel: [ 4743.454283] R13: 0000000000000200 R14= : 0000000000000000 R15: 0000000000001200 Dec 21 18:25:56 smilla kernel: [ 4743.454283] FS: 00007f4ed3929740(000= 0) GS:ffff88022fc00000(0000) knlGS:0000000000000000 Dec 21 18:25:56 smilla kernel: [ 4743.454283] CS: 0010 DS: 0000 ES: 00= 00 CR0: 0000000080050033 Dec 21 18:25:56 smilla kernel: [ 4743.454283] CR2: 0000000000000010 CR3= : 00000001014f4000 CR4: 00000000000006f0 Dec 21 18:25:56 smilla kernel: [ 4743.454283] DR0: 0000000000000000 DR1= : 0000000000000000 DR2: 0000000000000000 Dec 21 18:25:56 smilla kernel: [ 4743.454283] DR3: 0000000000000000 DR6= : 00000000ffff0ff0 DR7: 0000000000000400 Dec 21 18:25:56 smilla kernel: [ 4743.454283] Process blkid (pid: 16410= , threadinfo ffff8800be390000, task ffff88021dcddc00) Dec 21 18:25:56 smilla kernel: [ 4743.454283] Stack: Dec 21 18:25:56 smilla kernel: [ 4743.454283] 0000000000000001 0000000= 000000190 ffff88002bcabc00 ffff88021f0566c0 Dec 21 18:25:56 smilla kernel: [ 4743.454283] ffff8800be391948 fffffff= f814e5cb0 0000000000011210 ffff8800be391968 Dec 21 18:25:56 smilla kernel: [ 4743.454283] ffff88021dcddc00 0000000= 000000010 ffff8800be391928 ffffffff8111b0a5 Dec 21 18:25:56 smilla kernel: [ 4743.454283] Call Trace: Dec 21 18:25:56 smilla kernel: [ 4743.454283] [] md_= make_request+0xd0/0x200 Dec 21 18:25:56 smilla kernel: [ 4743.454283] [] ? m= empool_alloc_slab+0x15/0x20 Dec 21 18:25:56 smilla kernel: [ 4743.454283] [] gen= eric_make_request.part.50+0x74/0xb0 Dec 21 18:25:56 smilla kernel: [ 4743.454283] [] gen= eric_make_request+0x68/0x70 Dec 21 18:25:56 smilla kernel: [ 4743.454283] [] sub= mit_bio+0x85/0x110 ----- Opprinnelig melding ----- > All, >=20 > I have yet to find THE definitive word (via any search and forums) if > this is possible; expanding a raid0 set by adding drives. (Please, I > know there is no redundancy, and that is fine as these volumes are > really Amazon WS EBS volumes, so we only strip for speed improvements= ) >=20 >=20 >=20 > It seem this works: > more /proc/diskstats > mdadm --create /dev/md0 --chunk=3D256 --level=3D0 --raid-devices=3D2 > /dev/sd[b-c] > mdadm --detail /dev/md0 > mkfs.xfs /dev/md0 > mount /dev/md0 /mnt > vi /mnt/test > umount /dev/md0 >=20 > mdadm --grow /dev/md0 --level=3D0 --backup-file=3D/tmp/grow > --raid-devices=3D3 --add /dev/sdd <=3D changes to raid4 > mdadm --detail /dev/md0 > mdadm --grow /dev/md0 --level=3D0 --backup-file=3D/tmp/grow <=3D=3D c= hanges to > raid0 > mdadm --detail /dev/md0 >=20 > Can someone please confirm this correct and that raid0 can be > expanded, or is there a better means to expand an raid0 raid set? >=20 > Thanks! > -- > To unsubscribe from this list: send the line "unsubscribe linux-raid" > in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html --=20 Vennlige hilsener / Best regards roy -- Roy Sigurd Karlsbakk (+47) 98013356 roy@karlsbakk.net http://blogg.karlsbakk.net/ GPG Public key: http://karlsbakk.net/roysigurdkarlsbakk.pubkey.txt -- I all pedagogikk er det essensielt at pensum presenteres intelligibelt.= Det er et element=C3=A6rt imperativ for alle pedagoger =C3=A5 unng=C3=A5= eksessiv anvendelse av idiomer med xenotyp etymologi. I de fleste tilf= eller eksisterer adekvate og relevante synonymer p=C3=A5 norsk. -- To unsubscribe from this list: send the line "unsubscribe linux-raid" i= n the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html