[linux-lvm] Can someone _please_ tell me whats going on?

[linux-lvm] Can someone _please_ tell me whats going on?

[Bradley Alexander]

Hi all,

I posted last month (Please see message
<20011111194145.A715@sonsofthunder.yi.org>,
<20011112141948.P715@sonsofthunder.yi.org> regarding problems i was
having with lvm. I am working with a 1GHz Athlon machine, with 2 30GB
hard drives. One of the problems that I was seeing seek errors on the
hard drive. Everyone who responded mentioned that it was hardware
problems. 

Well, I took that Quantum drive out and put an almost brand new Maxtor
(I took great pleasure in whacking windows on it) identical to my first
drive (Maxtor Maxtor 53073U6). I did the pvcreate, vgcreate and
lvcreate, followed by mkreiserfs. When I rebooted, again, I saw 

hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hdc: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280,
sector=0
hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hdc: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280,
sector=0
hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hdc: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280,
sector=0
hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
hdc: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280,
sector=0
hdc: DMA disabled
hdd: DMA disabled
ide1: reset: success
hdc: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error
}
hdc: read_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280,
sector=0
hdc: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error
}
hdc: read_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280,
sector=0
hdc: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error
}

I know for a fact that this drive is good, as was the Quantum. However,
the Quantum has since stopped working. Obviously it is not a hardware
issue. Can someone please tell me why whenever I try to turn a second
drive into a logical volume, I get drive errors?

I would appreciate any suggestions on why this is happening and how to
fix it.


-- 
--Brad
============================================================================
Bradley M. Alexander, CISSP              |   Co-Chairman,
Beowulf System Admin/Security Specialist |    NoVALUG/DCLUG Security SIG
Debian/GNU Linux Developer		 |   storm [at] debian.org
                                         |   storm [at] tux.org
============================================================================
If Guns cause Crime, then Matches cause Arson.

Re: [linux-lvm] Can someone _please_ tell me whats going on?

[Friedrich Lobenstock]

On 16 Dec 2001 21:57, Bradley Alexander wrote:

> hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hdc: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280,
> sector=0
> hdc: DMA disabled
> hdd: DMA disabled
> ide1: reset: success

This looks like a IDE-cable problem to me. Checked the length?
Is it really a ATA66/100 cable? Do you use a drive cage?

-- 
MfG / Regards
Friedrich Lobenstock

Re: [linux-lvm] Can someone _please_ tell me whats going on?

[Bradley Alexander]

On Sun, 2001-12-16 at 22:13, Friedrich Lobenstock wrote:
> On 16 Dec 2001 21:57, Bradley Alexander wrote:
> 
> > hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> > hdc: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280,
> > sector=0
> > hdc: DMA disabled
> > hdd: DMA disabled
> > ide1: reset: success
> 
> This looks like a IDE-cable problem to me. Checked the length?
> Is it really a ATA66/100 cable? Do you use a drive cage?

Its a standard IDE cable. I have a 40-pin/40-wire cable and a
40-pin/80-wire (I'm assuming this is an ATA66/100 cable). Originally, i
had both drives on the ATA cable, but since I split them to different
channels, I had to use a standard IDE cable on the second channel.
However, I was getting the same errors when both were on the ATA cable
as well.

-- 
--Brad
============================================================================
Bradley M. Alexander, CISSP              |   Co-Chairman,
Beowulf System Admin/Security Specialist |    NoVALUG/DCLUG Security SIG
Debian/GNU Linux Developer		 |   storm [at] debian.org
                                         |   storm [at] tux.org
============================================================================
It's always a good idea to keep the pointy end going forward as
much as possible.
					--Rules of the Air, #21

Re: [linux-lvm] Can someone _please_ tell me whats going on?

[Bradley Alexander]

On Sun, 2001-12-16 at 22:43, Bradley Alexander wrote:
> 
> Its a standard IDE cable. I have a 40-pin/40-wire cable and a
> 40-pin/80-wire (I'm assuming this is an ATA66/100 cable). Originally, i
> had both drives on the ATA cable, but since I split them to different
> channels, I had to use a standard IDE cable on the second channel.
> However, I was getting the same errors when both were on the ATA cable
> as well.

I have noticed a couple more symptoms I thought I would forward.

1. The drive light stays illuminated when the machine is up, whether the
lv is mounted or not. Could this also be an indication of what killed
the first drive?

2. Doing normal actions on the volume group, in this case, vg01, causes
a segmentation fault. This includes vgchange, doing an fdisk on the
drive or anything else.

Anyone have any pointers?

thanks,
-- 
--Brad
============================================================================
Bradley M. Alexander, CISSP              |   Co-Chairman,
Beowulf System Admin/Security Specialist |    NoVALUG/DCLUG Security SIG
Debian/GNU Linux Developer		 |   storm [at] debian.org
                                         |   storm [at] tux.org
============================================================================
Ever notice when you blow in a dog's face he gets mad at you, but
when you take him in a car he sticks his head out the window?
						--George Carlin

Re: [linux-lvm] Can someone _please_ tell me whats going on?

[Wolfgang Weisselberg]

Hi, Bradley!

Bradley Alexander (storm@tux.org) wrote 41 lines:

> 1. The drive light stays illuminated when the machine is up, whether the
> lv is mounted or not. Could this also be an indication of what killed
> the first drive?

Uh-oH!  That doesn't look good... Bad cable?  Cable inserted
wrong?  IDE-Controller gone bad?

-Wolfgang

Re: [linux-lvm] Can someone _please_ tell me whats going on?

[Friedrich Lobenstock]

On 16 Dec 2001 22:43, Bradley Alexander wrote:

> Its a standard IDE cable. I have a 40-pin/40-wire cable and a
> 40-pin/80-wire (I'm assuming this is an ATA66/100 cable). Originally, i
> had both drives on the ATA cable, but since I split them to different
> channels, I had to use a standard IDE cable on the second channel.
> However, I was getting the same errors when both were on the ATA cable
> as well.
>

So please do yourself a favour and only use 80-wire cables nowadays.
That's why you are getting the
  hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
  hdc: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280, sector=0
messages.

-- 
MfG / Regards
Friedrich Lobenstock

Re: [linux-lvm] Can someone _please_ tell me whats going on?

[Bradley Alexander]

On Mon, 2001-12-17 at 07:20, Friedrich Lobenstock wrote:
> On 16 Dec 2001 22:43, Bradley Alexander wrote:
> 
> > Its a standard IDE cable. I have a 40-pin/40-wire cable and a
> > 40-pin/80-wire (I'm assuming this is an ATA66/100 cable). Originally, i
> > had both drives on the ATA cable, but since I split them to different
> > channels, I had to use a standard IDE cable on the second channel.
> > However, I was getting the same errors when both were on the ATA cable
> > as well.
> >
> 
> So please do yourself a favour and only use 80-wire cables nowadays.
> That's why you are getting the
>   hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
>   hdc: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280, sector=0
> messages.

I was getting the same type of errors even when both drives were on an
ATA cable. However, I will try to run out and get an ATA cable this
afternoon.

-- 
--Brad
============================================================================
Bradley M. Alexander, CISSP              |   Co-Chairman,
Beowulf System Admin/Security Specialist |    NoVALUG/DCLUG Security SIG
Debian/GNU Linux Developer		 |   storm [at] debian.org
                                         |   storm [at] tux.org
============================================================================
No, just _a_ Zaphod Beeblebrox. Didn't you hear? I come in six-packs."

Re: [linux-lvm] Can someone _please_ tell me whats going on?

[Bradley Alexander]

On Mon, 2001-12-17 at 12:38, Bradley Alexander wrote:
> On Mon, 2001-12-17 at 07:20, Friedrich Lobenstock wrote:
> > On 16 Dec 2001 22:43, Bradley Alexander wrote:
> > 
> > > Its a standard IDE cable. I have a 40-pin/40-wire cable and a
> > > 40-pin/80-wire (I'm assuming this is an ATA66/100 cable). Originally, i
> > > had both drives on the ATA cable, but since I split them to different
> > > channels, I had to use a standard IDE cable on the second channel.
> > > However, I was getting the same errors when both were on the ATA cable
> > > as well.
> > >
> > 
> > So please do yourself a favour and only use 80-wire cables nowadays.
> > That's why you are getting the
> >   hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> >   hdc: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280, sector=0
> > messages.
> 
> I was getting the same type of errors even when both drives were on an
> ATA cable. However, I will try to run out and get an ATA cable this
> afternoon.

I got a new cable, recreated the LV and copied everything over.
Rebooted, and was still getting the drive errors on boot, still getting
the drive light on solid, all of the problems I was having. To
summarize,

Swapped drives (after the first drive failed).
Swapped cables
Swapped controllers (was originally on the first controller with the   
   other drive, now on the second.)

The original drive (Quantum) was known good prior to starting to use
LVM, it has since failed. (It had, among other things, a Debian mirror
on it.) The replacememt drive is known good as well, no errors on the
drive itself.

vg00 on /dev/hda works fine, no problems encountered. vg01 (/dev/hdb)
has problems. It is allowing me to turn off the vg (vgchange -a n),
however, any operations done on that vg afterwards (including attempting
to try to turn it back on results in a segmentation fault. (see earlier
message in this thread for the exact trap information).

Can someone give me any advice on what to do to fix this problem?

-- 
--Brad
============================================================================
Bradley M. Alexander, CISSP              |   Co-Chairman,
Beowulf System Admin/Security Specialist |    NoVALUG/DCLUG Security SIG
Debian/GNU Linux Developer		 |   storm [at] debian.org
                                         |   storm [at] tux.org
============================================================================
If you push the stick forward, the houses get bigger. If you pull
the stick back, they get smaller. That is, unless you keep pulling
the stick all the way back, then they get bigger again.
					--Rules of the Air, #2

Re: [linux-lvm] Can someone _please_ tell me whats going on?

[Benjamin Scott]

On 17 Dec 2001, Bradley Alexander wrote:
> Swapped controllers (was originally on the first controller with the
>    other drive, now on the second.)

  FYI, both IDE ports on the motherboard use the same IDE controller chip,
so if it *is* the controller that is your problem (and given that your disk
activity light is stuck on, that may be), that will not fix it.

-- 
Ben Scott <bscott@ntisys.com>
| The opinions expressed in this message are those of the author and do not |
| necessarily represent the views or policy of any other person, entity or  |
| organization.  All information is provided without warranty of any kind.  |

Re: [linux-lvm] Can someone _please_ tell me whats going on?

[Lewis Bergman]

On Tuesday 18 December 2001 07:22 am, you wrote:
> On 17 Dec 2001, Bradley Alexander wrote:
> > Swapped controllers (was originally on the first controller with the
> >    other drive, now on the second.)
Seems strange since so many of us have not had any problems with LVM. Have 
you looked at the kernel notes for any clues to your particular hardware? You 
may find that some controllers/drives/PIO modes do not work together. Further 
tuning with hdparam may alleviate the problem.
As for myself. I wouldn't put much that is important on a Quantum as the 
quality seems to be fading. I have gone to IBM exclusively. Prices are good 
and I don't know of any better drive out there for ide or scsi.
-- 
Lewis Bergman
Texas Communications
4309 Maple St.
Abilene, TX 79602-8044
915-695-6962 ext 115

Re: [linux-lvm] Can someone _please_ tell me whats going on?

[Bradley Alexander]

I'm going to respond to both of these in the same post.

On Tue, 2001-12-18 at 08:22, Benjamin Scott wrote:
> On 17 Dec 2001, Bradley Alexander wrote:
> > Swapped controllers (was originally on the first controller with the
> >    other drive, now on the second.)
> 
>   FYI, both IDE ports on the motherboard use the same IDE controller chip,
> so if it *is* the controller that is your problem (and given that your disk
> activity light is stuck on, that may be), that will not fix it.

This is true, however, I now have (unlike at the point of my first
attempt to set this up) a vg on the first drive that is working
perfectly. I have seen no problems with it. The problems are only
occurring on the second drive. Also, for the record, I have created a
standard Linux partition via fdisk, made a reiserfs partition, and it is
happily schlepping data without hint of error or problem.

On Tue, 2001-12-18 at 08:50, Lewis Bergman wrote:
>
> Seems strange since so many of us have not had any problems with LVM.
> Have you looked at the kernel notes for any clues to your particular 
> hardware? You may find that some controllers/drives/PIO modes do not 
> work together. Further tuning with hdparam may alleviate the problem.
> As for myself. I wouldn't put much that is important on a Quantum as 
> the quality seems to be fading. I have gone to IBM exclusively. Prices
> are good and I don't know of any better drive out there for ide or 
> scsi.

The thing is that the drives are identical (and as I said, vg00 on
/dev/hda is operating without error), so if I was seeing problems, it
should stand to reason that I would see them on both drives.

I am looking at the kernel trap that LVM has been generating. It is
trapping on the vgchange process. I'm limping my way through it with
help.

-- 
--Brad
============================================================================
Bradley M. Alexander, CISSP              |   Co-Chairman,
Beowulf System Admin/Security Specialist |    NoVALUG/DCLUG Security SIG
Debian/GNU Linux Developer		 |   storm [at] debian.org
                                         |   storm [at] tux.org
============================================================================
Why are cigarettes sold at gas stations when smoking is prohibited?

Re: [linux-lvm] Can someone _please_ tell me whats going on?

[Steven Lembark]

> lvcreate, followed by mkreiserfs. When I rebooted, again, I saw
>
> hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hdc: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280,
> sector=0
> hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hdc: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280,
> sector=0
> hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hdc: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280,
> sector=0
> hdc: dma_intr: status=0x51 { DriveReady SeekComplete Error }
> hdc: dma_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280,
> sector=0
> hdc: DMA disabled
> hdd: DMA disabled
> ide1: reset: success
> hdc: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error
> }
> hdc: read_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280,
> sector=0
> hdc: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error
> }
> hdc: read_intr: error=0x10 { SectorIdNotFound }, LBAsect=211681280,
> sector=0
> hdc: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error
> }

(1) What happens if you use ext2 -- using one of the non-standard
    file systems for testing just brings in extra variables?


You could also have IDE controller or DMA problems. "Hardware"
doesn't just mean disk drives. If you have multiple controllers,
try switching the drives between them and see if that changes
the result.


--
Steven Lembark                               2930 W. Palmer
Workhorse Computing                       Chicago, IL 60647
                                            +1 800 762 1582

Re: [linux-lvm] Can someone _please_ tell me whats going on?

[Bradley Alexander]

On Sun, 2001-12-16 at 10:45, Steven Lembark wrote:

> (1) What happens if you use ext2 -- using one of the non-standard
>     file systems for testing just brings in extra variables?
 
I will try this, however, these errors occur prior to the filesystems
being initialized. Also, I decided to try to delete the PV. I deleted
the LV, but every time I tried to delete the VG, I get

[defiant /home/storm]# vgchange -a n vg01
vgchange -- volume group "vg01" successfully deactivated

[defiant /home/storm]# vgremove vg01
Segmentation fault

The logs show the following:

Dec 17 07:51:53 defiant kernel: Unable to handle kernel NULL pointer
dereference at virtual address 00000018
Dec 17 07:51:53 defiant kernel:  printing eip:
Dec 17 07:51:53 defiant kernel: c01589d9
Dec 17 07:51:53 defiant kernel: *pde = 00000000
Dec 17 07:51:53 defiant kernel: Oops: 0002
Dec 17 07:51:53 defiant kernel: CPU:    0
Dec 17 07:51:53 defiant kernel: EIP:    0010:[devfs_open+121/384]   
Tainted: P 
Dec 17 07:51:53 defiant kernel: EFLAGS: 00010286
Dec 17 07:51:53 defiant kernel: eax: 00000000   ebx: c1a65240   ecx:
dc144cc0   edx: c0267e00
Dec 17 07:51:53 defiant kernel: esi: 00000000   edi: c1a65244   ebp:
e7730100   esp: dade1f40
Dec 17 07:51:53 defiant kernel: ds: 0018   es: 0018   ss: 0018
Dec 17 07:51:53 defiant kernel: Process vgremove (pid: 956,
stackpage=dade1000)
Dec 17 07:51:53 defiant kernel: Stack: dc144cc0 e7730100 00000000
e7fea380 c025d7a0 c0130111 e7730100 dc144cc0 
Dec 17 07:51:53 defiant kernel:        bffffabc e34a1000 00000001
bffffabc c0130022 e7734940 e7fea380 00000001 
Dec 17 07:51:53 defiant kernel:        00000004 e34a1000 080525b8
e7734940 e7fea380 080525b8 bffffabc 00000000 
Dec 17 07:51:53 defiant kernel: Call Trace: [dentry_open+225/400]
[filp_open+82/96] [sys_open+54/160] [system_call+51/56] 
Dec 17 07:51:53 defiant kernel: 
Dec 17 07:51:53 defiant kernel: Code: 89 50 18 eb 2c 89 f6 8b 47 08 85
c0 74 1a 8b 10 85 d2 74 14 

 
> You could also have IDE controller or DMA problems. "Hardware"
> doesn't just mean disk drives. If you have multiple controllers,
> try switching the drives between them and see if that changes
> the result.

Originally, I had the drives on one IDE channel (hda and hdb) and the
DVD and CDRW on another (hdc and hdd). I now have drive 1 on hda, dvd on
hdb, drive 2 on hdc and cdrw on hdd. Same indications.

-- 
--Brad
============================================================================
Bradley M. Alexander, CISSP              |   Co-Chairman,
Beowulf System Admin/Security Specialist |    NoVALUG/DCLUG Security SIG
Debian/GNU Linux Developer		 |   storm [at] debian.org
                                         |   storm [at] tux.org
============================================================================
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