it seems like I keep running into a wall. The present raid array...well let me do an fdisk -l: ---------------------------------------- Disk /dev/hda: 250.0 GB, 250059350016 bytes 255 heads, 63 sectors/track, 30401 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System /dev/hda1 * 1 13 104391 fd Linux raid autodetect /dev/hda2 14 27651 222002235 fd Linux raid autodetect /dev/hda3 27652 30201 20482875 fd Linux raid autodetect /dev/hda4 30202 30401 1606500 5 Extended /dev/hda5 30202 30392 1534176 82 Linux swap
Disk /dev/hdc: 250.0 GB, 250059350016 bytes 255 heads, 63 sectors/track, 30401 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System /dev/hdc1 * 1 13 104391 fd Linux raid autodetect /dev/hdc2 14 27651 222002235 fd Linux raid autodetect /dev/hdc3 27652 30201 20482875 fd Linux raid autodetect /dev/hdc4 30202 30401 1606500 5 Extended /dev/hdc5 30202 30392 1534176 82 Linux swap
Disk /dev/md0: 106 MB, 106823680 bytes 2 heads, 4 sectors/track, 26080 cylinders Units = cylinders of 8 * 512 = 4096 bytes
Disk /dev/md0 doesn't contain a valid partition table
Disk /dev/md2: 20.9 GB, 20974338048 bytes 2 heads, 4 sectors/track, 5120688 cylinders Units = cylinders of 8 * 512 = 4096 bytes
Disk /dev/md2 doesn't contain a valid partition table
Disk /dev/md1: 227.3 GB, 227330162688 bytes 2 heads, 4 sectors/track, 55500528 cylinders Units = cylinders of 8 * 512 = 4096 bytes
Disk /dev/md1 doesn't contain a valid partition table
Disk /dev/sda: 250.0 GB, 250059350016 bytes 255 heads, 63 sectors/track, 30401 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System /dev/sda1 1 13 104391 fd Linux raid autodetect /dev/sda2 14 27651 222002235 fd Linux raid autodetect /dev/sda3 27652 30201 20482875 fd Linux raid autodetect /dev/sda4 30202 30401 1606500 5 Extended /dev/sda5 30202 30392 1534176 82 Linux swap
kk...the ide drives; hda and hdc...3 partitions 3 arrays md0 1 and 2 cat /proc/mdstat shows it's acting as it should.
when I boot the system up, it sees the drive connect thru usb and calls it sda (for what reason I don't know. I partitioned it to the same specs as the 2 ide drives as per the bottom of fdisk -l.
Now, trying to go by: http://www.redhat.com/docs/manuals/enterprise/RHEL-4-Manual/admin-guide/ 5.9.8.2 Creating RAID Arrays they state the following: Next, it is necessary to create the /etc/raidtab file. This file is responsible for the proper configuration of all RAID arrays on your system. The file format (which is documented in the raidtab(5) man page) is relatively straightforward. Here is an example /etc/raidtab entry for a RAID 1 array:
ummm...1.)I think this is part of the old raidtools thang...not mdadm but I am not sure. 2.) there is no man raidtab on this box which is a "everything" install..
so considering this..lol I'm lost.
kk, right now...I just formatted all the partitions and (I should have done this before ) I am checking for bad blocks. If it finds some I'll just redo it or whatever I have to do. At one point I was able to mdadm /dev/md0 1 & 2 --add sda1 2 &3 and cat /proc/mdstat would show sda1 2 &3 as in the arrays but then they would not populate. I can't remember but I think I gave them mount points but wait, I was having a hard time finding the proper syntax for that as well anyway, after rebooting...it was all gone.
this ain't nice
thx
John Rose
On Thu, 2006-09-28 at 18:16, rado wrote:
kk...the ide drives; hda and hdc...3 partitions 3 arrays md0 1 and 2 cat /proc/mdstat shows it's acting as it should.
Are you trying to add a 3rd partition to a raid created with only 2? I don't think that will work, although you could fail/remove an existing partition, then add the new one as a replacement.
when I boot the system up, it sees the drive connect thru usb and calls it sda (for what reason I don't know. I partitioned it to the same specs as the 2 ide drives as per the bottom of fdisk -l.
USB/firewire drives are assigned the next available scsi device name as detected - doesn't make much sense to me either...
ummm...1.)I think this is part of the old raidtools thang...not mdadm but I am not sure. 2.) there is no man raidtab on this box which is a "everything" install..
Raidtab isn't used any more - it's all mdadm commands.
kk, right now...I just formatted all the partitions and (I should have done this before ) I am checking for bad blocks. If it finds some I'll just redo it or whatever I have to do. At one point I was able to mdadm /dev/md0 1 & 2 --add sda1 2 &3 and cat /proc/mdstat would show sda1 2 &3 as in the arrays but then they would not populate.
It's only going to rebuild the number of devices you gave in the create command (or whever disk druid set up if you did it that way). I'm not sure if there is any way to change that later.
On Thu, 2006-09-28 at 20:36 -0500, Les Mikesell wrote:
On Thu, 2006-09-28 at 18:16, rado wrote:
kk...the ide drives; hda and hdc...3 partitions 3 arrays md0 1 and 2 cat /proc/mdstat shows it's acting as it should.
Are you trying to add a 3rd partition to a raid created with only 2? I don't think that will work, although you could fail/remove an existing partition, then add the new one as a replacement.
no, no...the existing arrays already have the "3rd" partition...I just want to add another complete drive, already partitioned exactly like the 2 that are already running. No extras there...just another drive.
when I boot the system up, it sees the drive connect thru usb and calls it sda (for what reason I don't know. I partitioned it to the same specs as the 2 ide drives as per the bottom of fdisk -l.
USB/firewire drives are assigned the next available scsi device name as detected - doesn't make much sense to me either...
ummm...1.)I think this is part of the old raidtools thang...not mdadm but I am not sure. 2.) there is no man raidtab on this box which is a "everything" install..
Raidtab isn't used any more - it's all mdadm commands.
exactly and there is a ton of stuff out there about raidtab grrrr and for 2 or 3 versions now it's all been mdadmin and rh still talks about raidtab in there docs...but I guess if that's the only thing wrong w/rh then we in pretty good shape.
kk, right now...I just formatted all the partitions and (I should have done this before ) I am checking for bad blocks. If it finds some I'll just redo it or whatever I have to do. At one point I was able to mdadm /dev/md0 1 & 2 --add sda1 2 &3 and cat /proc/mdstat would show sda1 2 &3 as in the arrays but then they would not populate.
It's only going to rebuild the number of devices you gave in the create command (or whever disk druid set up if you did it that way). I'm not sure if there is any way to change that later.
oh, c'mon Les, don't tell me that!!! geez! kk, I have not found it yet but I'm looking for some way to first add a spare to the existing raid sys...then go for it. hint hint lol. geez I dunno!
another thing that bothers me is that, just for the hell of it I started a new install. went to disk druid and the usb (sda) never was an option; like it never/doesn't exist. so of course I just backed out of it.
there has to be a way. I'm sure there is, you are doing the exact same thing I am trying to do!
thx Les,
John Rose
On Thu, 2006-09-28 at 21:54, rado wrote:
another thing that bothers me is that, just for the hell of it I started a new install. went to disk druid and the usb (sda) never was an option; like it never/doesn't exist. so of course I just backed out of it.
there has to be a way. I'm sure there is, you are doing the exact same thing I am trying to do!
My system is actually installed on /dev/sda (a scsi drive) and the 3-device raid was added later with:
mdadm --create /dev/md0 --level=1 --raid-devices=3 /dev/hdc1 /dev/hdd1 missing
The external drive becomes /dev/sdb when connected and can be added with mdadm --add /dev/md0 /dev/sdb1
I don't think you can do that in disk druid.
On Thu, 2006-09-28 at 23:24 -0500, Les Mikesell wrote:
My system is actually installed on /dev/sda (a scsi drive) and the 3-device raid was added later with:
kk now I am dealing w/2 ides on diff. channels and the removable usb recognized as sda
mdadm --create /dev/md0 --level=1 --raid-devices=3 /dev/hdc1 /dev/hdd1 missing
yes I remember you telling me this now...and something happened to that msg w/all the new vmware goin on here too...I dunno..
The external drive becomes /dev/sdb when connected and can be added with mdadm --add /dev/md0 /dev/sdb1
in my case would be sda1 2 3
I don't think you can do that in disk druid.
yes, this is one of your msgs that I lost...thx for this no gotta forget druid for this...
wondering if I should just dismantle the raid sys...and redo it as you did from the command line or...
seeing it's a brand new system...nothing really on it...just throw a fresh (choose auto for the partitions) then make the arrays.
I really would like to try to do what I'm trying to do just for the experience but lol I don't know who's gonna give out first me or raid...lol. It just seems hard to believe that this can't be done.
thx Les,
John Rose
so close yet so far or maybe that's just the way it is, I am dumb on it but gettin smarter.
the 2 ide drives diff channels of course hda and hdc each w/3 partitions hda1 2 3 hdc1 2 3.
I format hot add sda1 2 3 cat /proc/mdstat it's all there
...# mdadm /dev/md0 1 2 -f /dev/hdc1 2 3 (fail or fault hdc
md automatically picks up sda from spare status and writes to it.
it finishes mind you hdc is still under fail status and I reboot.
when the sys comes back up, it does not pick up sda anymore but instead just runs -U on all 3 partitions
I was under the impression that sda would be there either I missed a step somewhere or something strange going on
thx,
John Rose
On Fri, 2006-09-29 at 13:23 -0500, rado wrote:
so close yet so far or maybe that's just the way it is, I am dumb on it but gettin smarter.
the 2 ide drives diff channels of course hda and hdc each w/3 partitions hda1 2 3 hdc1 2 3.
I format hot add sda1 2 3 cat /proc/mdstat it's all there
Unless you created it with 3 members, the 3rd is added as a spare.
...# mdadm /dev/md0 1 2 -f /dev/hdc1 2 3 (fail or fault hdc
md automatically picks up sda from spare status and writes to it.
it finishes mind you hdc is still under fail status and I reboot.
when the sys comes back up, it does not pick up sda anymore but instead just runs -U on all 3 partitions
You can add /dev/hdc1 back if it didn't resync automatically.
I was under the impression that sda would be there either I missed a step somewhere or something strange going on
I don't think usb/firewire is detected early enough to be included in the raid assembly at bootup. That's why I set mine up to use 2 internal drives all the time but only periodically sync to the external drives that are then rotated offsite. And if you want to be sure everything is clean on the copy, you should stop any processes running on the mounted partition, and unmount the mount point momentarily while you fail the disk. If you don't, the contents would be the same as if the machine had crashed - probably still usable, but it could have problems.
Les, let's go back to private... w/this it's just you and I anyway k? oh...save the msg list bandwidth a bit.
thx John
Les Mikesell spake the following on 9/29/2006 11:48 AM:
On Fri, 2006-09-29 at 13:23 -0500, rado wrote:
so close yet so far or maybe that's just the way it is, I am dumb on it but gettin smarter.
the 2 ide drives diff channels of course hda and hdc each w/3 partitions hda1 2 3 hdc1 2 3.
I format hot add sda1 2 3 cat /proc/mdstat it's all there
Unless you created it with 3 members, the 3rd is added as a spare.
...# mdadm /dev/md0 1 2 -f /dev/hdc1 2 3 (fail or fault hdc
md automatically picks up sda from spare status and writes to it.
it finishes mind you hdc is still under fail status and I reboot.
when the sys comes back up, it does not pick up sda anymore but instead just runs -U on all 3 partitions
You can add /dev/hdc1 back if it didn't resync automatically.
I was under the impression that sda would be there either I missed a step somewhere or something strange going on
I don't think usb/firewire is detected early enough to be included in the raid assembly at bootup. That's why I set mine up to use 2 internal drives all the time but only periodically sync to the external drives that are then rotated offsite. And if you want to be sure everything is clean on the copy, you should stop any processes running on the mounted partition, and unmount the mount point momentarily while you fail the disk. If you don't, the contents would be the same as if the machine had crashed - probably still usable, but it could have problems.
Maybe I am too far down the message tree to get this, but I don't see how taking one drive from a raid 5 array off site does any good. One drive on a mirror gives you a backup. One drive in a raid5 just gives you a used hard drive with not much possibility to recover anything. But feel free to slap some sense into me if I am too far off. I could see if you had a raid 5 array, and then mirrored that to a raid 1 with the usb drive opposite the "entire" raid5 array you would have a viable backup. But you could just as easily mount the drive and rsync to it to get the same effect.
On Mon, 2006-10-02 at 13:10 -0700, Scott Silva wrote:
I format hot add sda1 2 3 cat /proc/mdstat it's all there
Unless you created it with 3 members, the 3rd is added as a spare.
Maybe I am too far down the message tree to get this, but I don't see how taking one drive from a raid 5 array off site does any good. One drive on a mirror gives you a backup. One drive in a raid5 just gives you a used hard drive with not much possibility to recover anything. But feel free to slap some sense into me if I am too far off. I could see if you had a raid 5 array, and then mirrored that to a raid 1 with the usb drive opposite the "entire" raid5 array you would have a viable backup. But you could just as easily mount the drive and rsync to it to get the same effect.
It probably wasn't clear from that conversation but it was about RAID1. The number of mirrors isn't limited to 2, so you can create a 3 drive array with one missing. This gives you 2 full-time mirrors plus the ability to add an external drive periodically, let the sync complete, then remove and rotate it offsite. It is best to unmount it before the fail/remove step, but it can run during the sync and you don't have to shut down.
On Mon, 2006-10-02 at 17:21 -0500, Les Mikesell wrote:
On Mon, 2006-10-02 at 13:10 -0700, Scott Silva wrote:
I format hot add sda1 2 3 cat /proc/mdstat it's all there
Unless you created it with 3 members, the 3rd is added as a spare.
Maybe I am too far down the message tree to get this, but I don't see how taking one drive from a raid 5 array off site does any good. One drive on a mirror gives you a backup. One drive in a raid5 just gives you a used hard drive with not much possibility to recover anything. But feel free to slap some sense into me if I am too far off. I could see if you had a raid 5 array, and then mirrored that to a raid 1 with the usb drive opposite the "entire" raid5 array you would have a viable backup. But you could just as easily mount the drive and rsync to it to get the same effect.
It probably wasn't clear from that conversation but it was about RAID1. The number of mirrors isn't limited to 2, so you can create a 3 drive array with one missing. This gives you 2 full-time mirrors plus the ability to add an external drive periodically, let the sync complete, then remove and rotate it offsite. It is best to unmount it before the fail/remove step, but it can run during the sync and you don't have to shut down.
...back to square one, Les...that other deal went south...I tried all day just would not let me add hda1 back into md0 soooo I'm back to this again lol...gonna make it work this time hope hope...and I will...prolly might do better after some sleep lol.
jr
On Mon, 2006-10-02 at 19:26, rado wrote:
...back to square one, Les...that other deal went south...I tried all day just would not let me add hda1 back into md0 soooo I'm back to this again lol...gonna make it work this time hope hope...and I will...prolly might do better after some sleep lol.
What error message did you get?
On Mon, 2006-10-02 at 20:10 -0500, Les Mikesell wrote:
On Mon, 2006-10-02 at 19:26, rado wrote:
...back to square one, Les...that other deal went south...I tried all day just would not let me add hda1 back into md0 soooo I'm back to this again lol...gonna make it work this time hope hope...and I will...prolly might do better after some sleep lol.
What error message did you get?
tryin to think...oh like one was ...it was busy(which it wasn't) then dmesg had a couple more...it really appeared to me like the kernel was flaking out some about it or something...really not sure but finally I said the hell w/this...and this aft...started to install and still ain't got it...I just finally wiped off the disk...I doubt it but wow when it rains it pours! I want to get this system at least up and running fore I go to bed...pulled all nighter again last nite...course that's steriods do that shit to me.
oh, I just rollin w/the punches w/this stuff, I not even gettin excited bout it...thx God huh?
lol
John
On Mon, 2006-10-02 at 13:10 -0700, Scott Silva wrote:
Les Mikesell spake the following on 9/29/2006 11:48 AM:
On Fri, 2006-09-29 at 13:23 -0500, rado wrote:
so close yet so far or maybe that's just the way it is, I am dumb on it but gettin smarter.
the 2 ide drives diff channels of course hda and hdc each w/3 partitions hda1 2 3 hdc1 2 3.
I format hot add sda1 2 3 cat /proc/mdstat it's all there
Unless you created it with 3 members, the 3rd is added as a spare.
...# mdadm /dev/md0 1 2 -f /dev/hdc1 2 3 (fail or fault hdc
md automatically picks up sda from spare status and writes to it.
it finishes mind you hdc is still under fail status and I reboot.
when the sys comes back up, it does not pick up sda anymore but instead just runs -U on all 3 partitions
You can add /dev/hdc1 back if it didn't resync automatically.
I was under the impression that sda would be there either I missed a step somewhere or something strange going on
I don't think usb/firewire is detected early enough to be included in the raid assembly at bootup. That's why I set mine up to use 2 internal drives all the time but only periodically sync to the external drives that are then rotated offsite. And if you want to be sure everything is clean on the copy, you should stop any processes running on the mounted partition, and unmount the mount point momentarily while you fail the disk. If you don't, the contents would be the same as if the machine had crashed - probably still usable, but it could have problems.
Maybe I am too far down the message tree to get this, but I don't see how taking one drive from a raid 5 array off site does any good. One drive on a mirror gives you a backup. One drive in a raid5 just gives you a used hard drive with not much possibility to recover anything.
We talking raid 1 Scott. It actually serves a duel purpose. It's part of a backup scheme plus it's a bootable disk image
But feel free to slap some sense into me if I am too far off. I could see if you had a raid 5 array, and then mirrored that to a raid 1 with the usb drive opposite the "entire" raid5 array you would have a viable backup. But you could just as easily mount the drive and rsync to it to get the same effect.
-
Les Mikesell -
rado -
Scott Silva