Re: Root RAID and unmounting /boot
On Tue, Oct 26, 1999 at 09:07:34AM -0600, Marcos Lopez wrote: Egon Eckert wrote: syslogd 378 root5w REG9,18548 192396 /var/log/boot.log klogd 389 root2r REG8,1 191102 12 /boot/System.map-2.2.12-20 Is it safe to kill these? These are loggers, so I guess nothing terrible would happen. But I wouldn't kill them anyway.. I would fix the problem, not make it worse. Also i would be quite grateful if someone could explain to me why I must unmount /boot inorder for the lilo -r /mnt/newroot to work? I don't unmount anything before running lilo. Okay tired performing that function without doing the umount. I tried doing it without coping the files in /boot over to /mnt/newroot/boot but it couldn't find /boot/boot.b. So i copied over all the files in /boot to /mnt/newroot/boot. Then after running "lilo -r /mnt/newroot" I get the following error: "Sorry, don't know how to handle device 0x0900" What does the above mean and how to i fix it? It's because /mnt/newroot/boot is not mounted on your boot partition, but is a regular subdirectory on your /mnt/newroot RAID. umount /boot, mount /mnt/newroot/boot. Then run LILO. -- : [EMAIL PROTECTED] : And I see the elder races, : :.: putrid forms of man: : Jakob Østergaard : See him rise and claim the earth, : :OZ9ABN : his downfall is at hand. : :.:{Konkhra}...:
RE: Build in degraded mode?
Mark Spencer wrote: as "failed-disk" as opposed to "raid-disk" in your /etc/raidtab file. The device you give it is not even actually accessed or written to; you could put /foo/bar and it will still build degraded without and error, if I'm not mistaken. mkraid says "unrecognized option "failed-disk" detected error line 13... You need a newer version of the raidtools package and the kernel patch. You can get the latest at ftp://ftp.us.kernel.org/pub/linux/daemons/raid/alpha this is the only way to build in degraded mode, so you will need to upgrade if you want to accomplish this. tom
AW: raid not auto detecting
Which kernel version are you using?? Have you patched it with the latest RAID patches? I'm running a stock rh 6.0 install. Where can i look to autodetect and start my raid arrays?
RE: Root RAID and unmounting /boot
my lilo.conf: disk=/dev/hda bios = 0x80 sectors = 63 heads = 128 cylinders = 779 partition = /dev/hda1 start = 63 boot=/dev/hda map=/boot/map install=/boot/boot.b prompt timeout=50 default=linux image=/boot/bzImage label=linux read-only root=/dev/md0 append="ether=0,0,0,0,eth1" Note the disk paragraph, My setup is hda1 as boot-disk (32mb) and the rest of my 2 hd's as raid0. Why exactly this is needed, i don't know. But without it, i also got the 0x0900 from lilo. In the context of booting from RAID, the disk geometry is only needed if the boot-loader is trying to read from or write to a file on a RAID-1 array that contains that partition. But in that case, the disk= and partition= lines would point to RAID arrays, not normal disks and partitions (see Harald Nordgård-Hansen's message to the list of 22 August on the "Booting Root RAID 1 Directly _Is_ Possible" thread). You may need this section in your lilo.conf to enable booting, but it has nothing to do with RAID, nor should it be a cure for the "Sorry, don't know how to handle device 0x0900" error, particularly as I believe your /dev/md0 is RAID-0, which cannot be read by lilo whatever parameters you pass to it. The lines in lilo.conf that could produce this error by pointing lilo at /dev/md0 are the "boot=", "map=", "install=", "loader=" or "image=" lines. In this case, it is clearly not the "boot=/dev/hda" line, as that is a straightforward installation of the boot-loader to the MBR on the first IDE disk. The "map=", "install=" and "image=" lines all point to files in /boot, and the absent "loader=" line defaults to /boot/chain.b as well, so if this is really the error message you are receiving, your problem is clearly something to do with /boot. You say that /dev/hda1 is your "boot-disk". Do you mean by this that /dev/hda1 is mounted on /boot? If so and you have all the normal files in /boot, I don't understand why you get the error message. If not, that is why you get the error. But even then, I still don't understand how adding disk geometry for /dev/hda has solved the problem. Something is not right with your description. Are you really saying that removing the disk geometry lines from your lilo.conf results in a "Sorry, don't know how to handle device 0x0900" error when you run lilo? The device number is particularly crucial. Could you try it again and check? Be careful that the only change you make it is to remove the geometry lines. And if you get an error, see if it is _exactly_ as stated above. Because as it stands, your solution doesn't make sense, as you yourself recognize. Cheers, Bruno Prior [EMAIL PROTECTED] -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Kelina Sent: 26 October 1999 18:04 To: Marcos Lopez; Egon Eckert Cc: [EMAIL PROTECTED] Subject: Re: Root RAID and unmounting /boot At 09:07 26/10/99 -0600, Marcos Lopez wrote: Egon Eckert wrote: syslogd 378 root5w REG9,18548 192396 /var/log/boot.log klogd 389 root2r REG8,1 191102 12 /boot/System.map-2.2.12-20 Is it safe to kill these? These are loggers, so I guess nothing terrible would happen. But I wouldn't kill them anyway.. I would fix the problem, not make it worse. Also i would be quite grateful if someone could explain to me why I must unmount /boot inorder for the lilo -r /mnt/newroot to work? I don't unmount anything before running lilo. Okay tired performing that function without doing the umount. I tried doing it without coping the files in /boot over to /mnt/newroot/boot but it couldn't find /boot/boot.b. So i copied over all the files in /boot to /mnt/newroot/boot. Then after running "lilo -r /mnt/newroot" I get the following error: "Sorry, don't know how to handle device 0x0900" What does the above mean and how to i fix it? Okay bear with me here... i'm doing this from memory and some of my own configs (if u have the time to search the archives, u'll find something abt this in august i think) my lilo.conf: disk=/dev/hda bios = 0x80 sectors = 63 heads = 128 cylinders = 779 partition = /dev/hda1 start = 63 boot=/dev/hda map=/boot/map install=/boot/boot.b prompt timeout=50 default=linux image=/boot/bzImage label=linux read-only root=/dev/md0 append="ether=0,0,0,0,eth1" Note the disk paragraph, My setup is hda1 as boot-disk (32mb) and the rest of my 2 hd's as raid0. Why exactly this is needed, i don't know. But without it, i also got the 0x0900 from lilo. The sectors, heads and cylinder's line are taken directly from fdisk -l /dev/hda. The start as far as i remember, is the same as the sector, since it's on the beginning of the disk. (correct me if i'm
RE: Root RAID and unmounting /boot
Marcos, My boot partition is on sda1 and that is on that partition. What do you mean by this? What is on that partition? Is /dev/sda1 mounted on /boot with all the normal /boot files on it? If so, your setup sounds fine. Assuming the setup is OK, you have 2 alternatives. Kill klogd, as you suggest. But you presumably have this running for a reason. So if you don't want to do this, the other alternative would be to leave /dev/sda1 on /boot for now. Edit /mnt/newroot/etc/lilo.conf so that you have an image section which selects your root-RAID array as root (e.g. root=/dev/md?). Delete the files you copied to /mnt/newroot/boot (you want it to be an empty mount-point). Make sure /mnt/newroot/etc/fstab shows / on your root-RAID device and /boot still on /dev/sda1. Now run lilo -C /mnt/newroot/etc/lilo.conf. If you now reboot and select the RAID image section at the boot prompt, everything should be working. Note to Jakob: The advice to move the boot device to /mnt/newroot/boot and running lilo with the -r option is more confusing than it needs to be. Why not simply make sure that lilo.conf includes a RAID section as above (where root=/dev/md?)? You don't need the boot device to be on /mnt/newroot/boot. All the files which lilo references are in the same physical place, whether it is mounted on /boot or /mnt/newroot/boot. You could replace the section beginning "Now, unmount the current /boot filesystem" and ending with "complete with no errors" with a simple instruction to run lilo without parameters, as long as you insert an item before the filesystem is copied across to the effect of: * Edit /etc/lilo.conf to add an image section which points to the root-RAID array as the root device (i.e. use "root=/dev/md?" where ? is the device number of your root RAID array). (We do this before the filesystem is copied so that /etc/lilo.conf matches /mnt/newroot/etc/lilo.conf, so we don't even have to use -C with lilo to pass it the new lilo.conf address.) I think this would be a good change to make as the unnecessary complication of moving the boot device seems to be tripping a lot of people up. Cheers, Bruno Prior [EMAIL PROTECTED] -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Marcos Lopez Sent: 25 October 1999 19:09 To: [EMAIL PROTECTED] Subject: Root RAID and unmounting /boot I am attempting to get my system to run RAID 1 for /, /usr, /var, /home I have moved over all except for / and it works fine. After reading the howto "Root file system on RAID. by - Jakob OEstergaard" I have decided to take the first approach, unfortunately I can not umount \boot. even after performing a umount -f \boot. I also performed a lsof | grep /boot to see what was being used and it returned the following. syslogd 378 root5w REG9,18548 192396 /var/log/boot.log klogd 389 root2r REG8,1 191102 12 /boot/System.map-2.2.12-20 Is it safe to kill these? Also i would be quite grateful if someone could explain to me why I must unmount /boot inorder for the lilo -r /mnt/newroot to work? My boot partition is on sda1 and that is on that partition. I actually had to copy my /boot to my /mnt/newroot/boot else there would be no files there. An explanation here would really be grand. What happens when i run lilo -r /mnt/newroot? How will this affect the current /boot as it is never touched as far as I can see.. Hehe, i am kinda confused :-) thanks -marcos
RE: Root RAID and unmounting /boot
There seems to be quite a lot of confusion about root-RAID out there. to /mnt/newroot/boot. Then after running "lilo -r /mnt/newroot" I get the following error: "Sorry, don't know how to handle device 0x0900" What does the above mean and how to i fix it? It does mean that lilo doesn't know anything about md device. Change the line containing "root=" in lilo conf to "root=0x900". This way lilo tells the kernel properly where from mount the root fs. This shouldn't be necessary. The root= line in /etc/lilo.conf is simply a pointer to a filesystem that is passed to the init scripts. It doesn't require reading of that filesystem, so lilo shouldn't have a problem with it pointing to a RAID array. I have root=/dev/md? in my machines with no problems. If lilo is reporting "Sorry, don't know how to handle device 0x0900", that means it is being asked to read from or write to /dev/md0. The lines in your lilo.conf that might require this include the "boot=", "install=", "map=", "loader=" and "image=" lines. If any of these lines points to something on /dev/md0, or if any of them is missing and the default (mostly in /boot) is on /dev/md0, and lilo cannot handle /dev/md0, then you will get this message. You can get round it by: (a) Getting the latest patched lilo from RedHat or applying the lilo.raid1 patch and rebuilding it yourself (if /dev/md0 is RAID-1) (b) Providing lilo with the geometry of one of the devices in the array (again if /dev/md0 is RAID-1) (c) Using a slightly-adapted grub instead of lilo (again if /dev/md0 is RAID-1) (d) Making sure the files to which these lines point are not on a software-RAID array. Check the mailing list archives for instructions on these methods. Yes. Put everything needed to run lilo to a non-raided partition. Almost. But not exactly. Put everything needed to *boot from* lilo on a non-raided partition. lilo.conf doesn't have to be on a non-raided partition itself. It isn't read during the boot process. It is read when installing lilo. At this time, your system is fully operational, so lilo can easily read it from a RAID array. The best advice would be to leave it in /etc, where it should be by default. Nor do you need /dev or any of its devices on a non-RAIDed partition. In fact, all you need on a non-RAIDed partition is the kernel image, the system map, boot.b and associated files. In other words, the files that live in /boot by default. So the simplest strategy is to stick with the default filesystem structure, and have /boot on its own tiny partition. Then you can use a bog standard /etc/lilo.conf, and run lilo straight without the need to pass it parameters. Cheers, Bruno Prior [EMAIL PROTECTED] -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Egon Eckert Sent: 26 October 1999 18:18 To: Marcos Lopez Cc: [EMAIL PROTECTED] Subject: Re: Root RAID and unmounting /boot to /mnt/newroot/boot. Then after running "lilo -r /mnt/newroot" I get the following error: "Sorry, don't know how to handle device 0x0900" What does the above mean and how to i fix it? It does mean that lilo doesn't know anything about md device. Change the line containing "root=" in lilo conf to "root=0x900". This way lilo tells the kernel properly where from mount the root fs. it get hammered. I think i am close just have to figure out this lilo thing. Yes. Put everything needed to run lilo to a non-raided partition. For example, this is my 'ls -lR /mnt/hda7' : boot: total 23 -rw-r--r-- 1 root root 512 Jul 28 16:09 boot.0300 -rw-r--r-- 1 root root 4540 Feb 2 1999 boot.b -rw-r--r-- 1 root root 612 Feb 2 1999 chain.b -rw--- 1 root root14336 Jul 28 16:09 map -rw-r--r-- 1 root root 620 Feb 2 1999 os2_d.b dev: total 0 brw-rw 1 root disk 3, 0 Jul 21 1998 hda brw-rw 1 root disk 3, 7 Jul 21 1998 hda7 lilo: total 933 -rw-r--r-- 1 root root 176 Jul 28 14:43 lilo.conf -rw-r--r-- 1 root root 451591 Jul 28 15:49 zImage -rw-r--r-- 1 root root 494994 Feb 4 1999 zImage-bak Then run lilo like this: archiv:/home/egon# lilo -r /mnt/hda7 -C /lilo/lilo.conf Added Linux-smp * Added Linux-bak archiv:/home/egon# Here's my lilo.conf: boot=/dev/hda root=0x900 install=/boot/boot.b map=/boot/mapvga=ext delay=30 image=/lilo/zImage label=Linux-smp image=/lilo/zImage-bak label=Linux-bak read-only Simple. Egon
Re: raid not auto detecting
On Wed, Oct 27, 1999 at 01:38:05PM +0200, Gordon Booth wrote: stock rh 6.0 already includes raid patches fyi it uses kernel 2.2.5 the correct questions are: have you changed the partition type to 0xfd? have you created an initrd with mkinitrd --with raid1 ... Which kernel version are you using?? Have you patched it with the latest RAID patches? I'm running a stock rh 6.0 install. Where can i look to autodetect and start my raid arrays? -- Luca Berra -- [EMAIL PROTECTED] Communications Media Services S.r.l.
Re: autorun
I forgot to set the partition type to 0xfd. Philipp Krause, SPTH-Projekt [EMAIL PROTECTED] http://www.spth.de
RE: raid not auto detecting
After setting up raid, the box is unable to autodetect the raid 1 array. doing a cat /proc/mdstat shows that the array is not running however when I type raidstart /dev/md0, then do a cat /proc/mdstat, everything is fine I'm running a stock rh 6.0 install. Where can i look to autodetect and start my raid arrays? Firstly, I take it this explains your previous problem with fsck failing on /dev/md0? fsck won't be able to do much if the array hasn't been initialized. This sounds like the classic RedHat 6.* auto-recognition problem. If you check dmesg ("dmesg | less"), is there a line that reads something like: kmod: failed to exec /sbin/modprobe -s -k md-personality-4, errno = 2 If so, this means that neither your initrd nor your kernel include RAID support. By default, RedHat 6.* comes with modular RAID support, which means that RAID arrays cannot be started until the filesystem which contains the modules (in /lib/modules) has been mounted. This occurs after the kernel tries to auto-recognize the arrays, so auto-recognition fails for lack of RAID support at that time. This explains why you can start it later with raidstart - /lib/modules is now available. To solve the problem, you can either rebuild the kernel to include RAID support built-in (as opposed to as a module), or you can add raid support to your initrd. To do this, run "mkinitrd --with=raid1 /boot/initrd 2.2.5-22", substituting your desired initrd name for /boot/initrd and the correct kernel version for 2.2.5-22. Make sure /etc/lilo.conf points to the initrd you just created in the "initrd=" line(s). Run lilo. RAID support should now be available during auto-recognition. The other thing to check is that you have marked the partitions which make up this array as type "fd" using fdisk. This tells the kernel that the partitions are part of an array. Without it, the partitions will be ignored during the auto-recognition phase. Make sure that your RAID is umounted and raidstopped before you run fdisk on the partitions. Cheers, Bruno Prior [EMAIL PROTECTED] -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Jason Wong Sent: 26 October 1999 22:05 To: [EMAIL PROTECTED] Subject: raid not auto detecting After setting up raid, the box is unable to autodetect the raid 1 array. doing a cat /proc/mdstat shows that the array is not running however when I type raidstart /dev/md0, then do a cat /proc/mdstat, everything is fine I'm running a stock rh 6.0 install. Where can i look to autodetect and start my raid arrays? TIA for any help.
Probable software RAID vs hardware RAID conflict -out of memory kernel crash
Dear Oh Great Raid Masters!! We're running RAID on two linux boxes. Both use software RAID, and one also has a hardware RAID (Mylex DAC960/DAC1100). Our kernel is 2.2.11 patched with raid0145-19990824-2.2.11. The box that only runs software RAID has no problems at all, whereas the box that also has hardware RAID crashes every day with "out of memory" and we have to reboot it. It's our department server, so it's making everyone unhappy. :( It turns out the used memory slowly increases until all the processes are swapped out and (we're guessing) the kernel itself runs out of memory. Could it be that the software RAID boot patch we applied doesn't work well with the hardware RAID code, that it introduces a memory leak in the hardware RAID code in the kernel? Looking at the code this seems probable since the patch replaces statically allocated arrays with dynamically allocated arrays (for some buffers) in code that could well be common between the two implementations. The software RAID patch probably adds the correct memory management functionality to the software RAID code, but maybe not the hardware RAID code. I'm not a kernel hackers, so this is the best bug report I can provide. Can you help me; do you have any pointers? Best regards, Jey "WorldSecure Server ogilvy.com" made the following annotations on 10/27/99 10:09:11 -- Privileged/Confidential Information may be contained in this message. If you are not the addressee indicated in this message (or responsible for delivery of the message to such person), you may not copy or deliver this message to anyone. In such case, you should destroy this message and kindly notify the sender by reply email. Please advise immediately if you or your employer does not consent to email for messages of this kind. Opinions, conclusions and other information in this message that do not relate to the official business of the Ogilvy Group shall be understood as neither given nor endorsed by it. ==
RE: autorun
Hello, I've been able to set up raid 1 for my /usr /var and /usr/local partitions on RH6.0 I've been lurking in the list for a while and I keep seeing references to setting the partition type to 0xfd for root-RAID. I feel dumb asking the question 'cuz the answer must be staring me in the face: How do you set the partition type to 0xfd? I don't see any reference to "fd" when I list the Hex codes (fdisk v2.9n) TIA Rich Fernandez -- From: [EMAIL PROTECTED][SMTP:[EMAIL PROTECTED]] Sent: Wednesday, October 27, 1999 8:30 AM To: LINUX-RAID-Mailingliste Subject: Re: autorun I forgot to set the partition type to 0xfd. Philipp Krause, SPTH-Projekt [EMAIL PROTECTED] http://www.spth.de
RE: Root RAID and unmounting /boot
On Wed, 27 Oct 1999, Bruno Prior wrote: BP (a) Getting the latest patched lilo from RedHat or applying the lilo.raid1 patch BP and rebuilding it yourself (if /dev/md0 is RAID-1) BP (b) Providing lilo with the geometry of one of the devices in the array (again BP if /dev/md0 is RAID-1) BP (c) Using a slightly-adapted grub instead of lilo (again if /dev/md0 is RAID-1) BP (d) Making sure the files to which these lines point are not on a software-RAID BP array. Just a note: I setup root RAID1 over the weekend on my RH61 box. The configuration file is really simple, you run lilo as normal, it writes the boot block to all disks in the array and you're done: boot=/dev/md0 map=/boot/map install=/boot/boot.b timeout=50 default=linux image=/boot/vmlinuz label=linux read-only root=/dev/md0 -- Randomly Generated Tagline: "How should I know if it works? That's what beta testers are for. I only coded it." (Attributed to Linus Torvalds, somewhere in a posting)
RE: Root RAID and unmounting /boot
In the context of booting from RAID, the disk geometry is only needed if the boot-loader is trying to read from or write to a file on a RAID-1 array that contains that partition. But in that case, the disk= and partition= lines would point to RAID arrays, not normal disks and partitions (see Harald Nordgård-Hansen's message to the list of 22 August on the "Booting Root RAID 1 Directly _Is_ Possible" thread). You may need this section in your lilo.conf to enable booting, but it has nothing to do with RAID, nor should it be a cure for the "Sorry, don't know how to handle device 0x0900" error, particularly as I believe your /dev/md0 is RAID-0, which cannot be read by lilo whatever parameters you pass to it. The lines in lilo.conf that could produce this error by pointing lilo at /dev/md0 are the "boot=", "map=", "install=", "loader=" or "image=" lines. In this case, it is clearly not the "boot=/dev/hda" line, as that is a straightforward installation of the boot-loader to the MBR on the first IDE disk. The "map=", "install=" and "image=" lines all point to files in /boot, and the absent "loader=" line defaults to /boot/chain.b as well, so if this is really the error message you are receiving, your problem is clearly something to do with /boot. You say that /dev/hda1 is your "boot-disk". Do you mean by this that /dev/hda1 is mounted on /boot? If so and you have all the normal files in /boot, I don't understand why you get the error message. If not, that is why you get the error. But even then, I still don't understand how adding disk geometry for /dev/hda has solved the problem. Something is not right with your description. Are you really saying that removing the disk geometry lines from your lilo.conf results in a "Sorry, don't know how to handle device 0x0900" error when you run lilo? The device number is particularly crucial. Could you try it again and check? Be careful that the only change you make it is to remove the geometry lines. And if you get an error, see if it is _exactly_ as stated above. Because as it stands, your solution doesn't make sense, as you yourself recognize. Hrm, it seems you are right... (figures:-) i just commented those lines out and it worked Must've been another mistake i made which i fixed when i added these lines to lilo.conf. Cuz i redid my whole conf file back then... Sorry for the misinformation :) Surge
RE: Root RAID and unmounting /boot
I've been following this thread as I research using software RAID1 on a production system I haven't read the new HOWTO yet (i plan to sometime today if i get time) but I dont think i understand the concept of bootable, root RAID1 so far. What I got so far is you have a small bootable partition which is not-raided, and then every other partition on that drive raided to like partitions on the mirror drive. The thing I TOTALLY don't get is if the first drive dies, how can you boot the 2nd drive (mirror) when you're still losing the small bootable partition (as it's still part of drive 1). Of course you can put this bootable partition on a seperate drive, but you still dont have the redundancy because THAT drive can die. Wouldn't it be easier to stick the kernel, lilo config, relevant boot info on a floppy and boot raid1 systems from that? perhaps i'm missing something..more likely that not :-) -Tony .-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-. Anthony J. Biacco Network Administrator/Engineer [EMAIL PROTECTED]Intergrafix Internet Services "Dream as if you'll live forever, live as if you'll die today" http://cygnus.ncohafmuta.comhttp://www.intergrafix.net .-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-._.-. On Wed, 27 Oct 1999, Bruno Prior wrote: There seems to be quite a lot of confusion about root-RAID out there. to /mnt/newroot/boot. Then after running "lilo -r /mnt/newroot" I get the following error: "Sorry, don't know how to handle device 0x0900" What does the above mean and how to i fix it? It does mean that lilo doesn't know anything about md device. Change the line containing "root=" in lilo conf to "root=0x900". This way lilo tells the kernel properly where from mount the root fs. This shouldn't be necessary. The root= line in /etc/lilo.conf is simply a pointer to a filesystem that is passed to the init scripts. It doesn't require reading of that filesystem, so lilo shouldn't have a problem with it pointing to a RAID array. I have root=/dev/md? in my machines with no problems. If lilo is reporting "Sorry, don't know how to handle device 0x0900", that means it is being asked to read from or write to /dev/md0. The lines in your lilo.conf that might require this include the "boot=", "install=", "map=", "loader=" and "image=" lines. If any of these lines points to something on /dev/md0, or if any of them is missing and the default (mostly in /boot) is on /dev/md0, and lilo cannot handle /dev/md0, then you will get this message. You can get round it by: (a) Getting the latest patched lilo from RedHat or applying the lilo.raid1 patch and rebuilding it yourself (if /dev/md0 is RAID-1) (b) Providing lilo with the geometry of one of the devices in the array (again if /dev/md0 is RAID-1) (c) Using a slightly-adapted grub instead of lilo (again if /dev/md0 is RAID-1) (d) Making sure the files to which these lines point are not on a software-RAID array. Check the mailing list archives for instructions on these methods. Yes. Put everything needed to run lilo to a non-raided partition. Almost. But not exactly. Put everything needed to *boot from* lilo on a non-raided partition. lilo.conf doesn't have to be on a non-raided partition itself. It isn't read during the boot process. It is read when installing lilo. At this time, your system is fully operational, so lilo can easily read it from a RAID array. The best advice would be to leave it in /etc, where it should be by default. Nor do you need /dev or any of its devices on a non-RAIDed partition. In fact, all you need on a non-RAIDed partition is the kernel image, the system map, boot.b and associated files. In other words, the files that live in /boot by default. So the simplest strategy is to stick with the default filesystem structure, and have /boot on its own tiny partition. Then you can use a bog standard /etc/lilo.conf, and run lilo straight without the need to pass it parameters. Cheers, Bruno Prior [EMAIL PROTECTED] -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Egon Eckert Sent: 26 October 1999 18:18 To: Marcos Lopez Cc: [EMAIL PROTECTED] Subject: Re: Root RAID and unmounting /boot to /mnt/newroot/boot. Then after running "lilo -r /mnt/newroot" I get the following error: "Sorry, don't know how to handle device 0x0900" What does the above mean and how to i fix it? It does mean that lilo doesn't know anything about md device. Change the line containing "root=" in lilo conf to "root=0x900". This way lilo tells the kernel properly where from mount the root fs. it get hammered. I think i am close just have to figure out this lilo thing. Yes. Put everything needed to run lilo to a non-raided partition. For example, this is my 'ls -lR /mnt/hda7' :
RE: DPT Linux RAID.
On Oct 26, 12:27pm, [EMAIL PROTECTED] wrote: } Subject: RE: DPT Linux RAID. On Tue, 26 Oct 1999, G.W. Wettstein wrote: CONFIG_SCSI_EATA_DMA The instability is especially profound in an SMP environment. Under any kind of load there will be crashes and hangs. Red Hat defaults to using CONFIG_SCSI_EATA_DMA if you have a PM2144UW. I have a client with 2 servers using PM2144UW's with CONFIG_SCSI_EATA_DMA that have been rock stable. Neither is SMP. One is a mail/web/dns server. The other is backup mail/dns and squid. That is an interesting datapoint. The PM3334 in a dual-PII simply would not hold up under load on our main IMAP server with the EATA_DMA driver. Our news server (dual PII-300) was also running with one of these and we saw problems. The problem seem to be compounded when we had SMC Etherpower-II (EPIC) cards in the machines. I had a lot of respect for SMC but these cards have no place in a production environment from our experience. The issue is probably pretty much moot for us at this point. The DPT cards, at least the 3334 we have, simply don't have the I/O performance that we need as our load scales. We have been pretty happy with the DAC960 cards although we had to turn off tagged queing in order to keep the drives on-line. We are moving toward fibre-channel and outboard RAID controllers to implement the SAN that we are deploying for our Linux server farms. Given the excellent luck that we have had with the software RAID code for Linux I probably see a diminishing future for hardware RAID controller cards in most of our servers. We are using software RAID1 to mirror root, var and swap and than deploying the service filesystems on the RAID5 composite volumes provided by the fibre-channel controllers. Jon Lewis *[EMAIL PROTECTED]*| Spammers will be winnuked or Thanks again for the note, have a pleasant remainder of the week. Greg }-- End of excerpt from [EMAIL PROTECTED] As always, Dr. G.W. Wettstein Enjellic Systems Development - Specializing 4206 N. 19th Ave.in information infra-structure solutions. Fargo, ND 58102 WWW: http://www.enjellic.com Phone: 701-281-1686 EMAIL: [EMAIL PROTECTED] -- "MS can classify NT however they like. Calling a pig a bird still doesn't get you flying ham, however." -- Steven N. Hirsch
Re: Probable software RAID vs hardware RAID conflict - out ofmemory kernel crash
2.2.11 has a memory leak in its tcp code. 2.2.12 is friendly, and I'm waiting to see how people take to 2.2.13 before test driving. The 2.2.11 patch should apply to 2.2.12 and you can ignore the 2 rejected hunks in fs.h On Wed, 27 Oct 1999 [EMAIL PROTECTED] wrote: - - -Dear Oh Great Raid Masters!! - -We're running RAID on two linux boxes. Both use software RAID, and one also -has a hardware RAID (Mylex DAC960/DAC1100). Our kernel is 2.2.11 patched with -raid0145-19990824-2.2.11. - -The box that only runs software RAID has no problems at all, whereas the box -that also has hardware RAID crashes every day with "out of memory" and we have -to reboot it. It's our department server, so it's making everyone unhappy. :( - -It turns out the used memory slowly increases until all the processes are -swapped out and (we're guessing) the kernel itself runs out of memory. Could it -be that the software RAID boot patch we applied doesn't work well with the -hardware RAID code, that it introduces a memory leak in the hardware RAID code -in the kernel? Looking at the code this seems probable since the patch replaces -statically allocated arrays with dynamically allocated arrays (for some buffers) -in code that could well be common between the two implementations. The software -RAID patch probably adds the correct memory management functionality to the -software RAID code, but maybe not the hardware RAID code. - -I'm not a kernel hackers, so this is the best bug report I can provide. Can you -help me; do you have any pointers? - -Best regards, -Jey - - - -"WorldSecure Server ogilvy.com" made the following - annotations on 10/27/99 10:09:11 --- -Privileged/Confidential Information may be contained in this message. If you are not the addressee indicated in this message (or responsible for delivery of the message to such person), you may not copy or deliver this message to anyone. In such case, you should destroy this message and kindly notify the sender by reply email. Please advise immediately if you or your employer does not consent to email for messages of this kind. Opinions, conclusions and other information in this message that do not relate to the official business of the Ogilvy Group shall be understood as neither given nor endorsed by it. - -== -
Re: autorun
On 10/27/1999 10:07 -0400, Fernandez, Richard wrote: Hello, I've been able to set up raid 1 for my /usr /var and /usr/local partitions on RH6.0 I've been lurking in the list for a while and I keep seeing references to setting the partition type to 0xfd for root-RAID. I feel dumb asking the question 'cuz the answer must be staring me in the face: How do you set the partition type to 0xfd? I don't see any reference to "fd" when I list the Hex codes (fdisk v2.9n) To quote a popular shoe maker... "Just do it". Never mind that fd isn't in the list; when it asks for the new partition type, type in '0xfd'. tw -- +--+--+ | Tim Walberg | Phone: 847-782-2472 | | TERAbridge Technologies Corp | FAX: 847-623-1717 | | 1375 Tri-State Parkway | [EMAIL PROTECTED] | | Gurnee, IL 60031 | 800-SKY-TEL2 PIN 9353299 | +--+--+ PGP signature
RE: mkraid aborts, no info?
Mandrake doesn't have RAID support built into the kernel AFAIK. I was trying to do the same thing you're doing using Mandrake 6.0. Below is an e-mail I received from Mandrake... dear Richard Fernandez, you should recompile the kernel with raid support or use the RedHat compiled kernel which already has that. sincerely, -- Florin Grad (Technical Support Team) [EMAIL PROTECTED] -- From: Bill Carlson[SMTP:[EMAIL PROTECTED]] Sent: Wednesday, October 27, 1999 4:17 PM To: [EMAIL PROTECTED] Subject: mkraid aborts, no info? Hello, I'm trying to setup your basic RAID0 Software raid and running into some problems. Details: Mandrake 6.1,(not RedHat!) Kernel 2.2.13 and 2.2.12 2 SCSI drives, completely seperate from root, 4 GB each My test raidtab (/etc/raidtab.test) raiddev /dev/md0 raid-level 0 nr-raid-disks 2 persistent-superblock 1 chunk-size 8 device /dev/sdb1 raid-disk 0 device /dev/sdc1 raid-disk 1 When I run mkraid I get the following: [root@washu /root]# mkraid -c /etc/raidtab.test /dev/md0 handling MD device /dev/md0 analyzing super-block disk 0: /dev/sdb1, 4192933kB, raid superblock at 4192832kB disk 1: /dev/sdc1, 4192933kB, raid superblock at 4192832kB mkraid: aborted, see the syslog and /proc/mdstat for potential clues. Output from /proc/mdstat: [root@washu /root]# chkraid Personalities : [2 raid0] read_ahead not set md0 : inactive md1 : inactive md2 : inactive md3 : inactive [root@washu /root]# Nothing of note in syslogd (*.* /dev/tty12) Any ideas on what I am doing wrong? I just joined the list, are there any searchable archives? Is a kernel patch still required for the 2.2.1x series? Thanks in advance, Bill Carlson Systems Programmer[EMAIL PROTECTED] | Opinions are mine, Virtual Hospital http://www.vh.org/| not my employer's. University of Iowa Hospitals and Clinics |
RE: mkraid aborts, no info?
On Wed, 27 Oct 1999, Fernandez, Richard wrote: Mandrake doesn't have RAID support built into the kernel AFAIK. I was trying to do the same thing you're doing using Mandrake 6.0. Below is an e-mail I received from Mandrake... dear Richard Fernandez, you should recompile the kernel with raid support or use the RedHat compiled kernel which already has that. sincerely, -- Florin Grad (Technical Support Team) [EMAIL PROTECTED] I'm thinking Florin means the kernel is not compiled with support by default. From the info on raidtools: This package includes the tools you need to set up and maintain a software RAID device under Linux. It only works with Linux 2.2 kernels and later, or 2.0 kernel specifically patched with newer raid support. To me that implies a 2.2.x kernel does not need a patch. On Mandrake 6.1, the required RAID modules were already in place after installation. Bill Carlson Systems Programmer[EMAIL PROTECTED]| Opinions are mine, Virtual Hospital http://www.vh.org/| not my employer's. University of Iowa Hospitals and Clinics|
RE: mkraid aborts, no info?
-- From: Bill Carlson[SMTP:[EMAIL PROTECTED]] Sent: Wednesday, October 27, 1999 4:30 PM To: Fernandez, Richard Cc: [EMAIL PROTECTED] Subject: RE: mkraid aborts, no info? On Wed, 27 Oct 1999, Fernandez, Richard wrote: Mandrake doesn't have RAID support built into the kernel AFAIK. I was trying to do the same thing you're doing using Mandrake 6.0. Below is an e-mail I received from Mandrake... dear Richard Fernandez, you should recompile the kernel with raid support or use the RedHat compiled kernel which already has that. sincerely, -- Florin Grad (Technical Support Team) [EMAIL PROTECTED] I'm thinking Florin means the kernel is not compiled with support by default. That's how I read it too. I re-compiled the kernel with support for RAID. Then I was able to "mkraid /dev/md0", but recieved "mkraid: aborted..." same as you. My /proc/mdstat showed the same as yours. Honestly, I never figured out how to fix it. We happened to have a copy of RH6 on hand and rather that "waste" any more time I installed it. Worked like a charm.
Re: mkraid aborts, no info?
On Wed, Oct 27, 1999 at 03:30:31PM -0500, Bill Carlson wrote: I'm thinking Florin means the kernel is not compiled with support by default. no. This package includes the tools you need to set up and maintain a software RAID device under Linux. It only works with Linux 2.2 kernels and later, or 2.0 kernel specifically patched with newer raid support. To me that implies a 2.2.x kernel does not need a patch. On Mandrake 6.1, the required RAID modules were already in place after installation. this is FALSE kernel 2.2.x does require a patch. get either a redhat kernel or a linux-2.2.13ac1 mandrake was unfortunate in that they included the raidtools package but did not patch the kernel to support raid. {to see if you have a patched kernel look for the "autodetecting raid" message during bootup} -- Luca Berra -- [EMAIL PROTECTED] Communications Media Services S.r.l.
Re: mkraid aborts, no info?
Hi Bill, You need to get the latest raid kernel patch (ignore the errors it gives... one hunk is included in the 2.2.12/2.2.13 kernel) and the latest raidtools (0.90). At 03:17 PM 10/27/99 -0500, Bill Carlson wrote: Hello, I'm trying to setup your basic RAID0 Software raid and running into some problems. Details: Mandrake 6.1,(not RedHat!) Kernel 2.2.13 and 2.2.12 2 SCSI drives, completely seperate from root, 4 GB each My test raidtab (/etc/raidtab.test) raiddev /dev/md0 raid-level 0 nr-raid-disks 2 persistent-superblock 1 chunk-size 8 device /dev/sdb1 raid-disk 0 device /dev/sdc1 raid-disk 1 When I run mkraid I get the following: [root@washu /root]# mkraid -c /etc/raidtab.test /dev/md0 handling MD device /dev/md0 analyzing super-block disk 0: /dev/sdb1, 4192933kB, raid superblock at 4192832kB disk 1: /dev/sdc1, 4192933kB, raid superblock at 4192832kB mkraid: aborted, see the syslog and /proc/mdstat for potential clues. Output from /proc/mdstat: [root@washu /root]# chkraid Personalities : [2 raid0] read_ahead not set md0 : inactive md1 : inactive md2 : inactive md3 : inactive [root@washu /root]# Nothing of note in syslogd (*.* /dev/tty12) Any ideas on what I am doing wrong? I just joined the list, are there any searchable archives? Is a kernel patch still required for the 2.2.1x series? Thanks in advance, Bill Carlson Systems Programmer[EMAIL PROTECTED]| Opinions are mine, Virtual Hospital http://www.vh.org/| not my employer's. University of Iowa Hospitals and Clinics| === David Cooley N5XMT Internet: [EMAIL PROTECTED] Packet: N5XMT@KQ4LO.#INT.NC.USA.NA T.A.P.R. Member #7068 Sponges grow in the ocean... Wonder how deep it would be if they didn't?! ===
Re: mkraid aborts, no info?
On Wed, 27 Oct 1999, David A. Cooley wrote: Hi Bill, You need to get the latest raid kernel patch (ignore the errors it gives... one hunk is included in the 2.2.12/2.2.13 kernel) and the latest raidtools (0.90). Ah, I see now. I'll try applying the patch to the 2.2.13 now. Thanks David, Luca. Bill Carlson Systems Programmer[EMAIL PROTECTED]| Opinions are mine, Virtual Hospital http://www.vh.org/| not my employer's. University of Iowa Hospitals and Clinics|
Re: raid problem
On Wed, Oct 27, 1999 at 11:09:46PM +0200, Mathieu ARNOLD wrote: I'm using a redhat 6.1 with a 2.3.21 kernel. well, i've gone back to the 2.2 series, i'm now using a 2.2.13 kernel, and well, it does exactly the same thing. does someone have a clue ? yup. neither 2.3.x nor 2.2.x do support new raid code, but you can find a patch for 2.2.x (while there is no patch for 2.3.x) i'd get ftp://ftp.fr.kernel.org/pub/linux/kernel/people/alan/2.2.13ac/patch-2.2.13ac1.gz other option is: ftp://ftp.fr.kernel.org/pub/linux/daemons/raid/alpha/raid0145-19990824-2.2.11.gz and fix the rejects by hand regards, Luca -- Luca Berra -- [EMAIL PROTECTED] Communications Media Services S.r.l.
raid0145 patch
Hey, The patch did the trick, just like it was supposed to. cd /usr/src/linux patch -p1 raid0145-19990824-2.2.11 There was the one error, which I ignored, as I was patching against 2.2.12. Does the same patch apply vs 2.2.13? I'm guessing that Mandrake's sources are what caused the errors that lead me to go with a fresh 2.2.12 source tree. Recompile, reboot and the magic messages started. :) 2 minutes later I had me an 8 GB array. Thanks a lot everyone! Bill Carlson Systems Programmer[EMAIL PROTECTED]| Opinions are mine, Virtual Hospital http://www.vh.org/| not my employer's. University of Iowa Hospitals and Clinics|
Re: raid problem
Luca Berra wrote: On Wed, Oct 27, 1999 at 11:09:46PM +0200, Mathieu ARNOLD wrote: I'm using a redhat 6.1 with a 2.3.21 kernel. ... well, i've gone back to the 2.2 series, i'm now using a 2.2.13 kernel, and well, it does exactly the same thing. does someone have a clue ? yup. neither 2.3.x nor 2.2.x do support new raid code, but you can find a patch for 2.2.x (while there is no patch for 2.3.x) i'd get ftp://ftp.fr.kernel.org/pub/linux/kernel/people/alan/2.2.13ac/patch-2.2.13ac1.gz other option is: ftp://ftp.fr.kernel.org/pub/linux/daemons/raid/alpha/raid0145-19990824-2.2.11.gz and fix the rejects by hand yep, i've read about it just after posting my msg :/ i now have : md0 : active raid5 hdh1[4] hdg1[3] hdf1[2] hde1[1] hdb1[0] 60066304 blocks level 5, 32k chunk, algorithm 2 [5/5] [U] resync=15% finish=163.4min and /dev/md0 56G 3.7G 53G 7% /opt1 i'm just happy now :) -- Cordialement Mathieu Arnold PGP key id : 0x2D92519F IRC : _mat ICQ 1827742 http://www.mat.cc/
raid problems solved - Thanks to all who helped
Well, after receiving numerous emails with suggestions and help, I've finally got this raid thing setup I just wanted to post and say thanks to all who helped. allan bruno mike surge luca gordon and anyone else that helped that I might have forgotten. This has been a really good experience and I must say that this is one of the best lists that I've ever subscribed to. Anyhow, here's a rundown of what I finally did to get a running system. Hopefully this will help in case another newbie decides to give this a whirl. 1. Installed stock RH 6.0 system using kernel 2.2.5-15 (This kernel has the necessary raid patch) 2. Setup up following partitions using fdisk during install process disk 0: /dev/sda116MB/boot 83 linux /dev/sda2128MB 82 linux swap /dev/sda3400MB / 83 linux /dev/sda45 extended /dev/sda5500MB fd unknown /dev/sda6500MB fd unknown /dev/sda7175MB fd unknown /dev/sda8175MB fd unknown disk 1: /dev/sdb116MB 83 linux /dev/sdb2128MB 82 linux swap /dev/sdb3400MB 83 linux /dev/sdb45 extended /dev/sdb5500MB fd unknown /dev/sdb6500MB fd unknown /dev/sdb7175MB fd unknown /dev/sdb8175MB fd unknown (to set the partition type to fd, type the command 't', select the partition you want to change and type 'fd' when prompted for the Hex code) 3. Finish installation process. 4. Reboot system. 5. Login as root 6. Install raidtools-0.90-3.i386.rpm from the distro CD 7. Setup /etc/raidtab per instructions from http://ostenfeld.dk/~jakob/Software-RAID.HOWTO/Software-RAID.HOWTO.txt for raid 1 8. run 'mkraid /dev/md0' 9. run 'mke2fs -b 4096 /dev/md0 10. run 'mkinitrd --with=raid1 /boot/initrd 2.2.5-15 11. edit /etc/lilo.conf - change line 'initrd=/boot/initrd-2.2.5-15.img' to 'initrd=/boot/initrd' 12. run 'lilo' 13. Reboot system 14. Login as root 15. run 'cat /proc/mdstat' and it should show an active raid setup Thanks again for all the help. Now to setup a bootable raid setup. :)
simple question
I am sure this is probably obvious but I wanna make sure:) I have a raid 1 setup using two ide disks. I am not booting from these disks. The RAID seems to be working GREAT so far:) I assume I should just put a normal entry in the fstab right? /dev/md0 /data ext2defaults1 2 fsck wont screw the raid array up right if its run against it? Thanks.. Mike -- Friends help you move. Real friends help you move bodies.
raid performance? good?
Is this good software raid performance for a Dual celeron 500 mhz 128 mb ram abit bp6 mb using the htp66 controller in dma 33 mode 2 x 18 gig ibm dma 33 drives Running kernel 2.2.13ac1, big ide patch (latest as of today), 2.2.14pre1 as well.. Its a raid 1 mirror with each disk alone on its own ide channel (one 3, one 4). OS is on a different drive and controller. ---Sequential Output ---Sequential InputRandom-- -Per Char- --Block--- -Rewrite-- -Per Char- --BlockSeeks--- MachineMB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU 400 7644 98.3 15585 23.1 7000 21.1 7861 92.5 17368 12.4 145.4 3.2 hdparm -Tt /dev/md0 /dev/md0: Timing buffer-cache reads: 64 MB in 0.85 seconds =75.29 MB/sec Timing buffered disk reads: 32 MB in 1.87 seconds =17.11 MB/sec I can understand the write performance but I would think the read performance would be better given that it should be reading from both halves of the mirror? I am using a 4096 block size on these drives if that helps. Basically I will be storing a lot of big files, database..etc. Any tuning suggestions? Basically looking for the MAXimum performance out of this thing. I am sure more ram will help which is my next purchase. Thanks, Mike -- Friends help you move. Real friends help you move bodies.
fsck and raid 1
I just rebooted my system for the first time automatically mounting up md0 through fstab. and I got a fsck failure on the md0 device. I made sure i let it totally sync before rebooting so I am not sure whats wrong. It was great until fsck ran against it. It seems that the parition table and the superblock dont agree on what the raid size is. Are there any known problems with fsck and raid v90 patch? fsck is from redhat 6.0. I am running kernel 2.2.13 After i created the raid device I mounted it up right away and i got these messages to the console right after the mount. md0: blocksize changed during read nr_blocks changed to 128 (blocksize 1024, j 264192, max_blocks 17619712) md0: blocksize changed during read nr_blocks changed to 32 (blocksize 4096, j 66048, max_blocks 4404928) the 4404928 is on of the numbers fsck complained about. It said the partition is actually smaller but the superblock says its larger (4404928). Any ideas? Mike -- Friends help you move. Real friends help you move bodies.
fsck and raid 1 more info
Here are the actual messages that fsck spits out when the system boots.. well basically the same. I am just running it manually here. [root@server1 /]# fsck /dev/md0 Parallelizing fsck version 1.14 (9-Jan-1999) e2fsck 1.14, 9-Jan-1999 for EXT2 FS 0.5b, 95/08/09 The filesystem size (according to the superblock) is 4404952 blocks The physical size of the device is 4404928 blocks Either the superblock or the partition table is likely to be corrupt! Aborty? And a copy of my /proc/mdstat [root@server1 /]# more /proc/mdstat Personalities : [linear] [raid0] [raid1] [raid5] read_ahead 1024 sectors md0 : active raid1 hdg1[1] hde1[0] 17619712 blocks [2/2] [UU] unused devices: none And a copy of the raid tab if this is useful.. [root@server1 /etc]# more raidtab raiddev /dev/md0 raid-level1 nr-raid-disks 2 nr-spare-disks0 chunk-size4 persistent-superblock 1 device/dev/hde1 raid-disk 0 device/dev/hdg1 raid-disk 1 Thanks for ANY info you can offer.. Mike - Previous email: I just rebooted my system for the first time automatically mounting up md0 through fstab. and I got a fsck failure on the md0 device. I made sure i let it totally sync before rebooting so I am not sure whats wrong. It was great until fsck ran against it. It seems that the parition table and the superblock dont agree on what the raid size is. Are there any known problems with fsck and raid v90 patch? fsck is from redhat 6.0. I am running kernel 2.2.13 After i created the raid device I mounted it up right away and i got these messages to the console right after the mount. md0: blocksize changed during read nr_blocks changed to 128 (blocksize 1024, j 264192, max_blocks 17619712) md0: blocksize changed during read nr_blocks changed to 32 (blocksize 4096, j 66048, max_blocks 4404928) the 4404928 is on of the numbers fsck complained about. It said the partition is actually smaller but the superblock says its larger (4404928). Any ideas? Mike -- Friends help you move. Real friends help you move bodies.
RE: Please Help - SCSI RAID device very slow
At 05:39 PM 10/27/1999 +0100, Matthew Clark wrote: Does anyone know how I stop the MegaRAID using the onboard INTEL SCSI Controller? (i960) It should be using my SYMBIOS card...??? It will only use the card it is physically connected to. Move the cable to the SYMBIOS card and it will be seen there. === David Cooley N5XMT Internet: [EMAIL PROTECTED] Packet: N5XMT@KQ4LO.#INT.NC.USA.NA T.A.P.R. Member #7068 We are Borg... Prepare to be assimilated! ===
RE: Please Help - SCSI RAID device very slow
Apparently the i960 is used to control the Symbios somehow...still - I doubt this explains the horrendus performance.. I am beginning to suspect a hardware problem elsewhere.. -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of David Cooley Sent: 27 October 1999 21:08 To: [EMAIL PROTECTED]; linux-scsi; linux-raid Subject: RE: Please Help - SCSI RAID device very slow At 05:39 PM 10/27/1999 +0100, Matthew Clark wrote: Does anyone know how I stop the MegaRAID using the onboard INTEL SCSI Controller? (i960) It should be using my SYMBIOS card...??? It will only use the card it is physically connected to. Move the cable to the SYMBIOS card and it will be seen there. === David Cooley N5XMT Internet: [EMAIL PROTECTED] Packet: N5XMT@KQ4LO.#INT.NC.USA.NA T.A.P.R. Member #7068 We are Borg... Prepare to be assimilated! === - To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to [EMAIL PROTECTED]
RE: Please Help - SCSI RAID device very slow
On Wed, 27 Oct 1999, Matthew Clark wrote: I do want to use the hardware RAID.. I think I have been looking at this for too long.. I now think the INTEL i960 IS the MegaRAID controller.. I didn't realise it was registered as a PCI device... How come it connects to the Symbios controller? I am very confused now!! Some days I think I'm perpetually confused, but hopefully this particular time, I'm the cluefull one. Does anyone know how I stop the MegaRAID using the onboard INTEL SCSI Controller? (i960) It should be using my SYMBIOS card...??? The AMI MegaRAIDs I've worked with (the 428) had an Intel i960 RISC processor integrated onto the megaraid card, along with several SymBIOS 8xx series SCSI controller chips. The i960 generates parity data and controls the RAID arrays, while the Symbios chips interface to the HDs under control of the i960. The newer MegaRAID cards can now "hijack" your other Symbios based controllers and make them part of the "collective" that is the megaraid. How slow is too slow? What numbers do you get from hdparm -tT? Looking at /proc/scsi/scsi, it appears that linux is talking directly to my MegaRAID controller.. Shouldn't it be talking to it via the SymBIOS controller? No, the MegaRAID is what it should be talking to. The Symbios controllers are now subsystems to the MegaRAID.
RE: Please Help - SCSI RAID device very slow
I do want to use the hardware RAID.. I think I have been looking at this for too long.. I now think the INTEL i960 IS the MegaRAID controller.. I didn't realise it was registered as a PCI device... How come it connects to the Symbios controller? I am very confused now!! -Original Message- From: casler, heather [mailto:[EMAIL PROTECTED]] Sent: 27 October 1999 18:54 To: '[EMAIL PROTECTED]' Subject: RE: Please Help - SCSI RAID device very slow If you don't want to use the MegaRaid at all, you might want to go into the system's BIOS and turn the card off in the BIOS. That was the only way I got the HP NetServer LH3 that I was working with to use the Symbios controller. Hope that helps. Heather -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] Sent: Wednesday, October 27, 1999 12:39 PM To: linux-scsi; linux-raid Subject: RE: Please Help - SCSI RAID device very slow Does anyone know how I stop the MegaRAID using the onboard INTEL SCSI Controller? (i960) It should be using my SYMBIOS card...??? -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Matthew Clark Sent: 27 October 1999 15:51 To: linux-scsi; linux-raid Subject: Please Help - SCSI RAID device very slow Looking at /proc/scsi/scsi, it appears that linux is talking directly to my MegaRAID controller.. Shouldn't it be talking to it via the SymBIOS controller? Host: scsi2 Channel: 01 Id: 00 Lun: 00 Vendor: MegaRAID Model: LD0 RAID5 26031R Rev: D Type: Direct-Access ANSI SCSI Revision: 02 Any help would be appreciated.. the disk array is running unacceptably slowly when writing - could this be the cause? Regards, Matthew Clark. -- NetDespatch Ltd - The Internet Call Centre. http://www.netdespatch.com - To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to [EMAIL PROTECTED]
RE: Please Help - SCSI RAID device very slow
Thanks for replying... The AMI MegaRAIDs I've worked with (the 428) had an Intel i960 RISC processor integrated onto the megaraid card, along with several SymBIOS 8xx series SCSI controller chips. The i960 generates parity data and controls the RAID arrays, while the Symbios chips interface to the HDs under control of the i960. This is what someone else has mentioned .. so the i960 is actually controlling the onboard Symbios controllers... The newer MegaRAID cards can now "hijack" your other Symbios based controllers and make them part of the "collective" that is the megaraid. Yikes - That sounds unplesant! Fortunately my other controller is an adaptec U2W (aic7xxx) How slow is too slow? What numbers do you get from hdparm -tT? Well .. hdparm -tT isn't "so" bad - here's the output #hdparm -tT /dev/sda /dev/sda: Timing buffer-cache reads: 64 MB in 0.61 seconds =104.92 MB/sec Timing buffered disk reads: 32 MB in 1.50 seconds =21.33 MB/sec (are 32 64 meg enough for these tests??) it's writing that's the BIG problem - check out this output from bonnie.. ---Sequential Output ---Sequential Input-- --Random-- -Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks--- MachineMB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU 1* 300 446 7.0 447 1.7 2440 4.9 4926 61.4 26823 24.6 1047.2 14.1 :-( No, the MegaRAID is what it should be talking to. The Symbios controllers are now subsystems to the MegaRAID. Right - well I'm glad you said that.. that was confusing the hell out of me! Both you and Gerard Roudier have confirmed that now.. I never heard of the i960 controlling controllers like that.. mind you I never claimed to be a hardware guru :-) Regards, Matthew Clark.
Re: Please Help - SCSI RAID device very slow
Does anyone know how I stop the MegaRAID using the onboard INTEL SCSI Controller? (i960) It should be using my SYMBIOS card...??? An i960 isnt a scsi controller. If you want to use the symbios directly load the ncr53c8xx driver and not the symbios driver. I'not totally sure if that will work but it is at least the right driver. That will basically cut out use of the megaraid totally
Re: Please Help - SCSI RAID device very slow
I do want to use the hardware RAID.. I think I have been looking at this for too long.. I now think the INTEL i960 IS the MegaRAID controller.. I didn't realise it was registered as a PCI device... How come it connects to the Symbios controller? Linux | [MegaRAID] ( Does raid stuff, drives scsi card) | [Symbios] ( SCSI card, does scsi not raid) | DISK
Please Help - SCSI RAID device very slow
Looking at /proc/scsi/scsi, it appears that linux is talking directly to my MegaRAID controller.. Shouldn't it be talking to it via the SymBIOS controller? Host: scsi2 Channel: 01 Id: 00 Lun: 00 Vendor: MegaRAID Model: LD0 RAID5 26031R Rev: D Type: Direct-Access ANSI SCSI Revision: 02 Any help would be appreciated.. the disk array is running unacceptably slowly when writing - could this be the cause? Regards, Matthew Clark. -- NetDespatch Ltd - The Internet Call Centre. http://www.netdespatch.com
RE: Please Help - SCSI RAID device very slow
Does anyone know how I stop the MegaRAID using the onboard INTEL SCSI Controller? (i960) It should be using my SYMBIOS card...??? -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Matthew Clark Sent: 27 October 1999 15:51 To: linux-scsi; linux-raid Subject: Please Help - SCSI RAID device very slow Looking at /proc/scsi/scsi, it appears that linux is talking directly to my MegaRAID controller.. Shouldn't it be talking to it via the SymBIOS controller? Host: scsi2 Channel: 01 Id: 00 Lun: 00 Vendor: MegaRAID Model: LD0 RAID5 26031R Rev: D Type: Direct-Access ANSI SCSI Revision: 02 Any help would be appreciated.. the disk array is running unacceptably slowly when writing - could this be the cause? Regards, Matthew Clark. -- NetDespatch Ltd - The Internet Call Centre. http://www.netdespatch.com