Ok well a d e f g h I are all RAID in disklabel I did newfs for all parts and raidctl -A root raid brings back; Raid0: Autoconfigure: Yes Raid0: Root: Yes
So that's seams present and correct. I am guessing I make it autoconfig then do newfs on the parts? But then I guess it doesn't matter which way round it happens does it? Well I did it after newfs and it displayed same output. And..ermmm yes :( missed that typo out, it should say i386 but you guessed that correctly :) Chris -----Original Message----- From: owner-m...@openbsd.org [mailto:owner-m...@openbsd.org] On Behalf Of Alexis de BRUYN Sent: 02 April 2009 21:44 To: Chris Harries Cc: misc@openbsd.org Subject: Re: FW: raidctl -vF component0 raid0 >From raidctl(8) : "-A root dev [snip] All components of the set must be of type RAID in the disklabel. [Snip] the RAID set must have its `a' partition (aka raid[0..n]a) set up. " >> raidctl -A root raid0 >> >> At this point, everything seams as expected >> >> Create radi0's partitions: >> Disklabel -E raid0 >> [snip] >> >> Create the new filesystems: >> For x in a d e f g h I; do newfs raid0${x}; done >> >> Seams to work fine. Now that your components are of type RAID, you must create your filesystems before making the RAID set auto-configurable: raidctl -A root raid0 and then: For x in a d e f g h I; do newfs raid0${x}; done I also noticed: >> Install 4.4 i386 on to sd0 [snip] >> cd /usr/src/sys/arch/sparc64/conf/ [snip] >> cp /usr/src/etc/etc.sparc64/MAKEDEV ./ Wrong copy/paste? Chris Harries a icrit : > Good call, did that, Still same problem, hangs at same place. > > All seams correct now: > > # disklabel -E raid0 > Initial label editor (enter '?' for help at any prompt) >> p > OpenBSD area: 0-1952459648; size: 1952459648; free: 719334272 > # size offset fstype [fsize bsize cpg] > a: 10485760 0 RAID > b: 4194304 10485760 swap > c: 1952459648 0 unused 0 0 > d: 104857600 14680064 RAID > e: 1048576000 119537664 RAID > f: 20971520 1168113664 RAID > g: 2097152 1189085184 RAID > h: 20971520 1191182336 RAID > i: 20971520 1212153856 RAID > > Noticed this in dmesg, not sure if at the half way point reboot though it's > anything to worry about. > > Kernelized RAIDframe activated > cd0(atapiscsi0:0:0): Check Condition (error 0x70) on opcode 0x0 > SENSE KEY: Not Ready > ASC/ASCQ: Medium Not Present > raid0 at root: (RAID Level 1) total number of sectors is 1952459648 (953349 > MB) as root > softraid0 at root > softraid0: sd0d can not read metadata version 1847620201, expected 3 > softraid0: sd1d can not read metadata version 8, expected 3 > softraid0: raid0a can not read metadata version 8, expected 3 > softraid0: raid0d can not read metadata version 8, expected 3 > softraid0: raid0e can not read metadata version 8, expected 3 > softraid0: raid0f can not read metadata version 8, expected 3 > softraid0: raid0g can not read metadata version 8, expected 3 > softraid0: raid0h can not read metadata version 8, expected 3 > softraid0: raid0i can not read metadata version 8, expected 3 > root on raid0a > filesystem type 19 not known.. assuming ffs > WARNING: / was not properly unmounted > swapmount: no device > raid0: Error re-writing parity! > > Chris > > -----Original Message----- > From: Alexis de BRUYN [mailto:ale...@de-bruyn.fr] > Sent: 02 April 2009 14:58 > To: Chris Harries > Cc: misc@openbsd.org > Subject: Re: FW: raidctl -vF component0 raid0 > >> Create radi0's partitions: >> Disklabel -E raid0 >> >> A: 5G / 4.2BSD >> B: 2048M swap >> D: 50G /var/mysql 4.2bsd >> E: 500G /var/vmail 4.2bsd >> F: 10G /var 4.2bsd >> G: 1G /tmp 4.2bsd >> H: 10G /usr 4.2bsd >> I: 10G /home 4.2bsd > > FS Type must be "RAID" instead of "4.2BSD". > > Chris Harries a icrit : >> All typoes checked. Now correct...finally, >> >> Apologies >> >> Chris >> >> -----Original Message----- >> From: Chris Harries [mailto:ch...@sharescope.co.uk] >> Sent: 02 April 2009 14:07 >> To: 'Alexis de BRUYN' >> Subject: RE: raidctl -vF component0 raid0 >> >> Afternoon, >> >> Well on an i386 system with 2 x 1 TB Seagate hard drives, I generally > follow >> this method from what I know and have learnt: >> >> Install 4.4 i386 on to sd0 >> Reboot >> >> Patch source tree to 4.4 PATCH: >> cd /usr/src >> mount /dev/cd0a /mnt >> tar -zxvpf /mnt/src.tar.gz -C /usr/src ./sys >> umount /mnt >> cvs -qd anon...@anoncvs.ca.openbsd.org:/cvs get -rOPENBSD_4_4 -P src >> >> Create new kernel with patches: >> cd /usr/src >> cvs -d anon...@anoncvs.uk.openbsd.org:/cvs up -Pd >> make -k cleandir >> rm -rf /usr/obj/* >> make obj >> >> >> cd /usr/src/etc/mtree >> install -c -o root -g wheel -m 660 special /etc/mtree >> install -c -o root -g wheel -m 444 4.4BSD.dist /etc/mtree >> mtree -qdef /etc/mtree/4.4BSD.dist -p / -u >> cd /usr/src/etc >> env DESTDIR=/ make distrib-dirs >> >> cd /usr/src/sys/arch/sparc64/conf/ >> config GENERIC >> cd ../compile/GENERIC >> make clean && make depend && make >> cp /bsd /bsd.old >> cp bsd /bsd >> chown root:wheel /bsd >> reboot >> >> cd /usr/src/ >> make build >> cd /dev >> cp /usr/src/etc/etc.sparc64/MAKEDEV ./ >> ./MAKEDEV all >> >> Sysmerge >> reboot >> >> Now we have a full patched system, add raidframe: >> cd /sys/arch/i386/conf >> cat >> GENERIC.RAID << EOF >> include "arch/'uname -m'/conf/GENERIC >> option RAID_AUTOCONFIG >> pseudo-device raid 4 >> EOF >> >> Re-create the kernel with patches and raidframe: >> config GENERIC.RAID >> cd ../compile/GENERIC.RAID >> make clean depend && make >> cp /bsd /bsd.noraid >> install -o root -g wheel -m 644 bsd / >> >> Test it boots ok, no problems. Now on to raidframe: >> fdisk -i sd1 >> (Yes to the MBR question) >> >> disklabel -E sd1 >> Create 512mb A: 4.2BSD >> Create *(rest of space) D: RAID >> >> Create new filesystem and mount: >> newfs sd1a >> mount /dev/sd1a /mnt >> cp /bsd /usr/mdec/boot /mnt >> /usr/mdec/installboot -v /mnt/boot /usr/mdec/biosboot sd1 >> umount /mnt >> >> Create raid0.conf >> cat >> /root/raid0.conf << EOF >> START array >> 1 2 0 >> >> START disks >> /dev/sd2d >> /dev/sd1d >> >> START layout >> 128 1 1 1 >> >> START queue >> fifo 100 >> EOF >> >> raidctl -C /root/raid0.conf raid0 >> raidctl -I 0904020 (the date) raid0 >> >> Check and add arrays: >> raidctl -s raid0 >> raidctl -A root raid0 >> >> At this point, everything seams as expected >> >> Create radi0's partitions: >> Disklabel -E raid0 >> >> A: 5G / 4.2BSD >> B: 2048M swap >> D: 50G /var/mysql 4.2bsd >> E: 500G /var/vmail 4.2bsd >> F: 10G /var 4.2bsd >> G: 1G /tmp 4.2bsd >> H: 10G /usr 4.2bsd >> I: 10G /home 4.2bsd >> >> Create the new filesystems: >> For x in a d e f g h I; do newfs raid0${x}; done >> >> Seams to work fine. >> >> Mount and copy RAID: >> mount /dev/raid0a /mnt >> cd /mnt >> mkdir usr tmp home var var/mysql var/vmail >> mount /dev/raid0d /mnt/var/mysql >> mount /dev/raid0e /mnt/var/vmail >> mount /dev/raid0f /mnt/var >> mount /dev/raid0g /mnt/tmp >> mount /dev/raid0h /mnt/usr >> mount /dev/raid0i /mnt/home >> >> Transfer raid: >> cd /mnt >> tar -Xcpf - / | tar -xvpf - >> rm /mnt/etc/fstab >> cat >> /mnt/etc/fstab << EOF >> /dev/raid0a / ffs rw 1 1 >> /dev/raid0d /var/mysql ffs rw 1 2 >> /dev/raid0e /var/vmail ffs rw 1 2 >> /dev/raid0f /var ffs rw 1 2 >> /dev/raid0g /tmp ffs rw 1 2 >> /dev/raid0h /usr ffs rw 1 2 >> /dev/raid0i /home ffs rw 1 2 >> EOF >> >> Umount partitions and reboot: >> umount /mnt/*; umount /mnt >> halt (reboot) >> >> boot> boot sd1a:/bsd to boot to 2nd hard disk >> >> Check, mirror, hot add, reconstruct: >> mount && uname -v && raidctl -s raid0 >> disklabel sd1 > /root/disklabel.sd1 >> disklabel -R sd0 /root/disklabel.sd1 >> raidctl -a /dev/sd0d raid0 >> raidctl -vF component0 raid0 >> >> And there is hangs, and displays >> >>> RECON: initiating reconstruction on row - col 0 -> spare at row 0 col 2. >>> Quiescence reached... >> How does this differ from what you do? >> >> Many thanks, again >> Chris >> >> -----Original Message----- >> From: Alexis de BRUYN [mailto:ale...@de-bruyn.fr] >> Sent: 02 April 2009 13:39 >> To: Chris Harries >> Cc: misc@openbsd.org >> Subject: Re: raidctl -vF component0 raid0 >> >> Hello Chris, >> >> Before setting up your mirror, I recommanded you to read "RAID options >> for OpenBSD" from the OpenBSD FAQ >> (http://www.openbsd.org/faq/faq14.html#RAID) and then the following >> manuals (which I did when I teached myself making a RAIDFRAME mirror on >> 4.2) : >> >> * raid(4), >> * raidctl(8), >> * newfs(8), >> * disklabel(8), >> * fdisk(8), >> * boot(8), >> * installboot(8), >> * dd(1), >> >> With the same steps, my configuration is working on 4.3 & 4.4 (amd64). >> >> You can also *precisely* describe your steps (commands and traces), and >> in this case, I could easely help you. >> >> Best regards, >> >> Chris Harries a icrit : >>> Thank you for your advice Alexis, I have now tried to do this using wd2d >> and >>> it does indeed make sense. I am still having problems however. Everything >>> seams to go fine, to what the 2 guides I am following suggest, but when >>> reconstructing the data is where I get stuck! >>> >>> When running raidctl -vF component0 raid0 I see >>> >>> RECON: initiating reconstruction on row - col 0 -> spare at row 0 col 2. >>> Quiescence reached... >>> >>> And that is where it stops, just sitting there. I am guessing when you do >>> the command it brings up a bar of how much it has reconstructed with > maybe >>> an ETA, but I don't see this, no hard drive light flashing. >>> >>> Befor that command I do >>> >>> disklabel wd1 > /root/disklabel.wd1 >>> disklabel -R wd0 /root/disklabel.wd1 >>> raidctl -a /dev/wd0b raid0 >>> >>> Which seams fine with me. Did you following a guide to teach your self >> this? >>> I have tried reading over man raidctl but it's now showing me anything >> more >>> then I know already and what I am not doing correct to cause this >>> reconstruction to just hang...? Any ideas >>> >>> Many Thanks >>> Chris >>> -----Original Message----- >>> From: Alexis de BRUYN [mailto:ale...@de-bruyn.fr] >>> Sent: 31 March 2009 12:33 >>> To: Chris Harries >>> Cc: misc@openbsd.org >>> Subject: Re: raidctl -vF component0 raid0 >>> >>>>> A: 144522 4.2BSD (this is the 64MB drive to boot off >>>>> B: 1953375480 RAID (this is the RAID data partition >>>>> C: 1953523055 UNUSED >>> Using 'b' (even 'c') is not a good idea for me too. >>> >>> Try on your second disk (mirror), before configuring RAID, with the two >>> following partitions: >>> >>> a: 512M 4.2BSD Boot partition >>> c: ----- unused Entire drive >>> d: * RAID Everything except boot kernel >>> >>> >>>>>>> START disks >>>>>>> /dev/wd2b # the fake device >>>>>>> /dev/wd1b >>>>>>> >>> And then: >>> >>> START disks >>> /dev/wd2d >>> /dev/wd1d >>> >>> It works for my several configurations all the times. >>> >>> Chris Harries a icrit : >>>> Thank you for your time. >>>> >>>> This I did find weird, wondering why on this guide, it is setting B to >>> RAID >>>> and not swap...on boot it does say it cannot find swap but this guide > did >>>> come recommended... >>>> >>>> It says >>>> >>>> A: 144522 4.2BSD (this is the 64MB drive to boot off >>>> B: 1953375480 RAID (this is the RAID data partition >>>> C: 1953523055 UNUSED >>>> >>>> I am guessing you meant wd0 and wd1, the guide suggested making wd2 as >> the >>>> fake device as I am creating the install on wd0, putting over to wd1 > then >>>> booting to wd1 and initializing wd0 again and create the raid, in a very >>> cut >>>> way to explain it >>>> >>>> Chris >>>> >>>> -----Original Message----- >>>> From: J.C. Roberts [mailto:list-...@designtools.org] >>>> Sent: 30 March 2009 13:16 >>>> To: Chris Harries >>>> Cc: misc@openbsd.org >>>> Subject: Re: raidctl -vF component0 raid0 >>>> >>>> On Mon, 30 Mar 2009 09:43:31 +0100 "Chris Harries" >>>> <ch...@sharescope.co.uk> wrote: >>>> >>>>> START disks >>>>> /dev/wd2b # the fake device >>>>> /dev/wd1b >>>>> >>>> The above looks weird. The 'b' partition is typically swap. >>>> >>>> What do the following commands tell you? >>>> >>>> $ sudo disklabel -n wd1 >>>> >>>> $ sudo disklabel -n wd2 >>>> >>>> >>> -- >>> Alexis de BRUYN >>> email : ale...@de-bruyn.fr >>> > > -- > Alexis de BRUYN > email : ale...@de-bruyn.fr > -- Alexis de BRUYN email : ale...@de-bruyn.fr