Re: [CentOS] (C8) root on mdraid
Mon, 16 Nov 2020 14:28:54 -0800 Gordon Messmer : > On 11/15/20 10:40 PM, Łukasz Posadowski wrote: > > > Sun, 15 Nov 2020 14:16:48 -0800 Gordon Messmer > > : > > > > > >> Use metadata version 1.2 instead of 0.9. > >> > > Thanks, I'll try that. I'm use to metadata 0.9, because GRUB have > > (had?) some issue with the newer ones. > > > If that doesn't work, and you need to use metadata 0.9, then check > /etc/default/grub and make sure that GRUB_CMDLINE_LINUX contains > "rd.md.uuid=". T.H.A.N.K Y.O.U. :) I was sure I had to pass uuid of the partition, not md0. It is working now. One last thing: since Linux Rescue CD does not recognize selinux, / mounted itself as read-only. Disabling SeLinux helped. I think I need to relabel all the files on a new host. So generally: - edit /etc/default/grub and grub2-mkconfig, - generate mdadm.conf (I think a copy of it resides inside initramfs), - fstab (root mounts anyway, so just for swap patition), - disable selinux, - generate initramfs with dracut. Big thanks to all of you. It was rough, but it's done. -- Łukasz Posadowski ___ CentOS mailing list CentOS@centos.org https://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] (C8) root on mdraid
Mon, 16 Nov 2020 10:29:17 + Tony Mountifield : > In article > <20201115123245.db62b8248e1f248afe028...@lukaszposadowski.pl>, Lukasz > Posadowski wrote: > > > > Hello everyone. > > > > I'm trying to install CentOS 8 with root and swap partitions on > > software raid. The plan is: > > - create md0 raid level 1 with 2 hard drives: /dev/sda and /dev/sdb, > > using Linux Rscue CD, > > - install CentOS 8 with Virtual Box on my laptop, > > - rsync CentOS 8 root partition on /dev/md0p1, > > - chroot in CentOS 8 root partition, > > - configure /etc/mdadm.conf, grub.cfg, initramfs, install > > bootloader on both sda and sdb drives. > > > > I think I can do first four of the above, but my CentOS installation > > acts strange after rebooting the server. It recognizes the raid, but > > boots randomly with root on /dev/sda1 (and recognizes raid > > with /dev/sdb disk), or with root on /dev/sdb1 (and recognizes raid > > with /dev/sda disk). When booting from Linux Rescue CD, the raid > > with two disk is recognized. > > I thought it was much more usual to partition both disks to give > sda1,2,3 and sdb1,2,3, and then create /dev/md0 from > sda1/sdb1, /dev/md1 from sda2/sdb3, and so on. > > That's the way I have always done it, and have never had any problems. > Never seen an attempt to partition an md device before. In that case, > how would the kernel and initrd be found in order to assemble the > RAID? It totaly works either way - raid on partitions and raid on disks. I'm use to mirroring whole disks. -- Łukasz Posadowski ___ CentOS mailing list CentOS@centos.org https://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] (C8) root on mdraid
On 11/15/20 10:40 PM, Łukasz Posadowski wrote: Sun, 15 Nov 2020 14:16:48 -0800 Gordon Messmer : Use metadata version 1.2 instead of 0.9. Thanks, I'll try that. I'm use to metadata 0.9, because GRUB have (had?) some issue with the newer ones. If that doesn't work, and you need to use metadata 0.9, then check /etc/default/grub and make sure that GRUB_CMDLINE_LINUX contains "rd.md.uuid=". ___ CentOS mailing list CentOS@centos.org https://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] (C8) root on mdraid
On Mon, Nov 16, 2020, 2:29 AM Tony Mountifield wrote: > > I thought it was much more usual to partition both disks to give sda1,2,3 > and sdb1,2,3, and then create /dev/md0 from sda1/sdb1, /dev/md1 from > sda2/sdb3, > and so on. > What I always did was to mdraid a single full disk partition then use lvm for any file systems.Boot disks did need a separate /boot partition. > ___ CentOS mailing list CentOS@centos.org https://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] (C8) root on mdraid
In article <20201115123245.db62b8248e1f248afe028...@lukaszposadowski.pl>, Lukasz Posadowski wrote: > > Hello everyone. > > I'm trying to install CentOS 8 with root and swap partitions on > software raid. The plan is: > - create md0 raid level 1 with 2 hard drives: /dev/sda and /dev/sdb, > using Linux Rscue CD, > - install CentOS 8 with Virtual Box on my laptop, > - rsync CentOS 8 root partition on /dev/md0p1, > - chroot in CentOS 8 root partition, > - configure /etc/mdadm.conf, grub.cfg, initramfs, install bootloader on > both sda and sdb drives. > > I think I can do first four of the above, but my CentOS installation > acts strange after rebooting the server. It recognizes the raid, but > boots randomly with root on /dev/sda1 (and recognizes raid > with /dev/sdb disk), or with root on /dev/sdb1 (and recognizes raid > with /dev/sda disk). When booting from Linux Rescue CD, the raid with > two disk is recognized. I thought it was much more usual to partition both disks to give sda1,2,3 and sdb1,2,3, and then create /dev/md0 from sda1/sdb1, /dev/md1 from sda2/sdb3, and so on. That's the way I have always done it, and have never had any problems. Never seen an attempt to partition an md device before. In that case, how would the kernel and initrd be found in order to assemble the RAID? Cheers Tony -- Tony Mountifield Work: t...@softins.co.uk - http://www.softins.co.uk Play: t...@mountifield.org - http://tony.mountifield.org ___ CentOS mailing list CentOS@centos.org https://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] (C8) root on mdraid
Hi, I am curious, what config your /etc/grub2.cfg looks like? does it have the correct root partition specified under "set root=" or in linux option or how it looks in boot loader during a start? Thanks On Mon, 16 Nov 2020 at 08:41, Łukasz Posadowski wrote: > Sun, 15 Nov 2020 14:16:48 -0800 > Gordon Messmer : > > > On 11/15/20 3:32 AM, Łukasz Posadowski wrote: > > > Do anyone can suggest what else I forgot to do? > > > > > > Use metadata version 1.2 instead of 0.9. > > > > You need for the filesystem to be not visible until after the RAID is > > assembled, and the easiest way to do that is to put the metadata at > > the beginning of the drive and the partition table inside the RAID > > volume. > > > > With metadata version 0.9, the partition table is technically inside > > the volume, but it's at the same offset that it would be for a disk > > with no RAID volume, so it can be recognized before the volume is > > assembled. > > Thanks, I'll try that. I'm use to metadata 0.9, because GRUB have > (had?) some issue with the newer ones. > > -- > Łukasz Posadowski > ___ > CentOS mailing list > CentOS@centos.org > https://lists.centos.org/mailman/listinfo/centos > -- Ruslanas Gžibovskis +370 6030 7030 ___ CentOS mailing list CentOS@centos.org https://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] (C8) root on mdraid
Sun, 15 Nov 2020 14:16:48 -0800 Gordon Messmer : > On 11/15/20 3:32 AM, Łukasz Posadowski wrote: > > Do anyone can suggest what else I forgot to do? > > > Use metadata version 1.2 instead of 0.9. > > You need for the filesystem to be not visible until after the RAID is > assembled, and the easiest way to do that is to put the metadata at > the beginning of the drive and the partition table inside the RAID > volume. > > With metadata version 0.9, the partition table is technically inside > the volume, but it's at the same offset that it would be for a disk > with no RAID volume, so it can be recognized before the volume is > assembled. Thanks, I'll try that. I'm use to metadata 0.9, because GRUB have (had?) some issue with the newer ones. -- Łukasz Posadowski ___ CentOS mailing list CentOS@centos.org https://lists.centos.org/mailman/listinfo/centos
Re: [CentOS] (C8) root on mdraid
On 11/15/20 3:32 AM, Łukasz Posadowski wrote: Do anyone can suggest what else I forgot to do? Use metadata version 1.2 instead of 0.9. You need for the filesystem to be not visible until after the RAID is assembled, and the easiest way to do that is to put the metadata at the beginning of the drive and the partition table inside the RAID volume. With metadata version 0.9, the partition table is technically inside the volume, but it's at the same offset that it would be for a disk with no RAID volume, so it can be recognized before the volume is assembled. ___ CentOS mailing list CentOS@centos.org https://lists.centos.org/mailman/listinfo/centos
[CentOS] (C8) root on mdraid
Hello everyone. I'm trying to install CentOS 8 with root and swap partitions on software raid. The plan is: - create md0 raid level 1 with 2 hard drives: /dev/sda and /dev/sdb, using Linux Rscue CD, - install CentOS 8 with Virtual Box on my laptop, - rsync CentOS 8 root partition on /dev/md0p1, - chroot in CentOS 8 root partition, - configure /etc/mdadm.conf, grub.cfg, initramfs, install bootloader on both sda and sdb drives. I think I can do first four of the above, but my CentOS installation acts strange after rebooting the server. It recognizes the raid, but boots randomly with root on /dev/sda1 (and recognizes raid with /dev/sdb disk), or with root on /dev/sdb1 (and recognizes raid with /dev/sda disk). When booting from Linux Rescue CD, the raid with two disk is recognized. Creating md0 wth Linux Rescue CD. # lsblk NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT loop0 7:00 819.5M 1 loop /run/archiso/sfs/airootfs sda 8:00 232.9G 0 disk sdb 8:16 0 232.9G 0 disk sr011:01 871M 0 rom # mdadm --create /dev/md0 --level=1 --raid-devices=2 --metadata=0.9 /dev/sda /dev/sdb mdadm: array /dev/md0 started. # mdadm --detail --scan --verbose ARRAY /dev/md0 level=raid1 num-devices=2 metadata=0.90 UUID=fef2f68b: 1a462c80:c44c77eb:7ee19756 devices=/dev/sda,/dev/sdb # cat /proc/mdstat Personalities : [raid1] md0 : active raid1 sdb[1] sda[0] 244198464 blocks [2/2] [UU] [>] resync = 1.9% (4682944/244198464) finish=52.2min speed=76356K/sec bitmap: 2/2 pages [8KB], 65536KB chunk unused devices: (1 hour later) # cat /proc/mdstat Personalities : [raid1] md0 : active raid1 sdb[1] sda[0] 244198464 blocks [2/2] [UU] bitmap: 0/2 pages [0KB], 65536KB chunk unused devices: # lsblk NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT loop0 7:00 819.5M 1 loop /run/archiso/sfs/airootfs sda 8:00 232.9G 0 disk └─md0 9:00 232.9G 0 raid1 sdb 8:16 0 232.9G 0 disk └─md0 9:00 232.9G 0 raid1 sr011:01 871M 0 rom - So far so good. Now partitions: # fdisk /dev/md0 Welcome to fdisk (util-linux 2.33.1). Changes will remain in memory only, until you decide to write them. Be careful before using the write command. Device does not contain a recognized partition table. Created a new DOS disklabel with disk identifier 0x64998720. Command (m for help): o Created a new DOS disklabel with disk identifier 0xfe784f57. Command (m for help): n Partition type p primary (0 primary, 0 extended, 4 free) e extended (container for logical partitions) Select (default p): p Partition number (1-4, default 1): First sector (2048-488396927, default 2048): Last sector, +/-sectors or +/-size{K,M,G,T,P} (2048-488396927, default 488396927): +220G Created a new partition 1 of type 'Linux' and of size 220 GiB. Command (m for help): n Partition type p primary (1 primary, 0 extended, 3 free) e extended (container for logical partitions) Select (default p): p Partition number (2-4, default 2): First sector (461375488-488396927, default 461375488): Last sector, +/-sectors or +/-size{K,M,G,T,P} (461375488-488396927, default 488396927): +4G Created a new partition 2 of type 'Linux' and of size 4 GiB. Command (m for help): w The partition table has been altered. Calling ioctl() to re-read partition table. Syncing disks. # lsblk NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT loop0 7:00 819.5M 1 loop /run/archiso/sfs/airootfs sda 8:00 232.9G 0 disk └─md0 9:00 232.9G 0 raid1 ├─md0p1 259:00 220G 0 part └─md0p2 259:10 4G 0 part sdb 8:16 0 232.9G 0 disk └─md0 9:00 232.9G 0 raid1 ├─md0p1 259:00 220G 0 part └─md0p2 259:10 4G 0 part sr011:01 871M 0 rom # mkfs -t ext4 /dev/md0p1 mke2fs 1.44.5 (15-Dec-2018) Creating filesystem with 57671680 4k blocks and 14417920 inodes Filesystem UUID: 511d0d07-4d55-4b6c-939e-d23ae2da1ce0 Superblock backups stored on blocks: 32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, 4096000, 7962624, 11239424, 2048, 23887872 Allocating group tables: done Writing inode tables: done Creating journal (262144 blocks): done Writing superblocks and filesystem accounting information: done # mkswap /dev/md0p2 Setting up swapspace version 1, size = 4 GiB (4294963200 bytes) no label, UUID=5667d007-0d3e-4b54-8a72-6ade52609adc -- Mount would be root partition: # mkdir /mnt/centos8 # mount -t ext4 /dev/md0p1 /mnt/centos8/ -- OK, I have what I need. Now log into virtualised CentOS 8 in Virtual Box. # rpm -q rsync rsync-3.1.3-7.el8.x86_64 # rpm -q mdadm mdadm-4.1-13.el8.x86_64 # rsync -vaxAXHSz --delete / 192.168.111.167:/mnt/centos8/ The authenticity of host '192