All of lore.kernel.org
 help / color / mirror / Atom feed
From: Sean Burford <sean.burford@adelaide.edu.au>
To: linux-lvm@sistina.com
Subject: Re: [linux-lvm] lvcreate core dump on large disks, vg name corrupted.
Date: Thu, 27 Sep 2001 20:30:19 +0930	[thread overview]
Message-ID: <3BB306C3.29ECD29F@adelaide.edu.au> (raw)
In-Reply-To: 3BB300E1.90809@st-peter.stw.uni-erlangen.de

Hi,

I have recompiled with a clean 2.4.9 kernel, using the following
patches:
tar xIf /usr/local/src/linux-2.4.9.tar.bz2
cat /usr/local/src/bonding-2.4.9-20010920 | patch -p1  (used for
etherchannel bonding, nothing to do with filesystems)
patch -p1 <
/usr/local/src/LVM/1.0.1-rc2/PATCHES/lvm-1.0.1-rc2-2.4.9.patch
patch -p1 <
/usr/local/src/LVM/1.0.1-rc2/PATCHES/linux-2.4.9-VFS-lock.patch
I'm also using the HP lpfcdd fiberchannel module to access the SAN
storage.

dmesg now reports:
  LVM version 1.0.1-rc2(30/08/2001) module loaded

I still get segfaults from lvcreate.

And have recreated the partitions as 2*20G, and still get segfaults with
the same symptoms (lv_create() passed junk in lv_name), so disk size
does not appear to be the key.

I might try it against a couple of loopback filesystems without lpfcdd
and bonding loaded, to see if it is a problem interacting with the
lpfcdd module.  I'll try creating filesystems on the SAN LUN's/disks
first.

svetljo wrote:
> 
> you can not use lvm-1.0.1rc2 tools with 0.9.1beta2 code in the kernel
> 
> try to patch the kernel with the lvm code from cvs
> 
> or befor configuring the lvm code change the kernel version in the top
> Makefile
> /usr/src/linux/Makefile  to 2.4.8-ac14 and after applying the patch back
> to 2.4.9-ac14
> this should be working with lvm-1.0.1rc2
> 
> or use lvm-1.0, the patch applies clean to -ac kernel
> 
> Sean Burford wrote:
> 
> >Hi,
> >
> >I am attempting to use the LVM tools under Linux 2.4.9-ac14 to set up
> >access to a SAN.  lvcreate is seg faulting because main() is passing a
> >corrupted vg name to lv_create().  I suspect this is due to the large
> >partitions that are being dealt with (4*100G).
> >
> >Using:
> >  LVM 1.0.1-rc2 userland tools
> >  The LVM kernel modules that came with the kernel.  I could not apply
> >the 1.0.x patches as they do not compile with the Alan Cox patches.  I
> >will see if the straight 2.4.9 kernel works with the AMI MegaRAID card
> >in the box soon, because if it does I will be able to patch the kernel
> >for LVM 1.0.1-rc2.
> >
> >Some related output from dmesg is:
> >SCSI device sdh: 209715200 512-byte hdwr sectors (107374 MB)
> > sdh: sdh1
> >SCSI device sdi: 209715200 512-byte hdwr sectors (107374 MB)
> > sdi: sdi1
> >SCSI device sdj: 209715200 512-byte hdwr sectors (107374 MB)
> > sdj: sdj1
> >SCSI device sdk: 209715200 512-byte hdwr sectors (107374 MB)
> > sdk: sdk1
> >LVM version LVM 0.9.1_beta7(ish)(20 June 2001) module loaded
> >
> >Following the LVM-Howto, I have configured the LVM as follows:
> >vgscan
> >pvcreate /dev/sdh1
> >pvcreate /dev/sdi1
> >pvcreate /dev/sdj1
> >pvcreate /dev/sdk1
> >vgcreate san_vg /dev/sd[hijk]1
> >lvcreate -v -L255000 -ndata_lv san_vg
> >
> >lvcreate segfaults:
> ># lvcreate -v -L255000 -ndata_lv san_vg
> >lvcreate -- checking volume group name "san_vg"
> >lvcreate -- checking volume group existence
> >lvcreate -- checking volume group activity
> >lvcreate -- checking logical volume path "/dev/san_vg/data_lv"
> >lvcreate -- checking stripe count
> >lvcreate -- checking stripe size
> >lvcreate -- locking logical volume manager
> >lvcreate -- getting volume group status from VGDA in kernel
> >lvcreate -- checking stripe size against volume group physical extent
> >size
> >lvcreate -- reading volume group data of "san_vg"
> >lvcreate -- checking logical volume maximum size
> >lvcreate -- checking volume group free space
> >lvcreate -- checking stripe count against physical volume count
> >lvcreate -- checking for maximum logical volume count
> >lvcreate -- setting up logical volume
> >lvcreate -- setting read ahead sectors
> >lvcreate -- creating logical volume VGDA in kernel
> >Segmentation fault (core dumped)
> >
> >I recompiled it with debugging, and got the following output:
> >
> >[root@backgammon tools]# gdb ./lvcreate
> >GNU gdb 5.0rh-5 Red Hat Linux 7.1
> >Copyright 2001 Free Software Foundation, Inc.
> >GDB is free software, covered by the GNU General Public License, and you
> >are
> >welcome to change it and/or distribute copies of it under certain
> >conditions.
> >Type "show copying" to see the conditions.
> >There is absolutely no warranty for GDB.  Type "show warranty" for
> >details.
> >This GDB was configured as "i386-redhat-linux"...
> >(gdb) run -v -L255000 -ndata_lv san_vg
> >Starting program: /usr/local/src/LVM/1.0.1-rc2/tools/./lvcreate -v
> >-L255000 -ndata_lv san_vg
> >lvcreate -- checking volume group name "san_vg"
> >lvcreate -- checking volume group existence
> >lvcreate -- checking volume group activity
> >lvcreate -- checking logical volume path "/dev/san_vg/data_lv"
> >lvcreate -- checking stripe count
> >lvcreate -- checking stripe size
> >lvcreate -- locking logical volume manager
> >lvcreate -- getting volume group status from VGDA in kernel
> >lvcreate -- checking stripe size against volume group physical extent
> >size
> >lvcreate -- reading volume group data of "san_vg"
> >lvcreate -- checking logical volume maximum size
> >lvcreate -- checking volume group free space
> >lvcreate -- checking stripe count against physical volume count
> >lvcreate -- checking for maximum logical volume count
> >lvcreate -- setting up logical volume
> >lvcreate -- setting read ahead sectors
> >lvcreate -- creating logical volume VGDA in kernel
> >
> >Program received signal SIGSEGV, Segmentation fault.
> >lv_create (vg=0x5, lv=0xbffffa14,
> >    lv_name=0xbffffa2c
> >"dûÿ¿\213ûÿ¿°ûÿ¿àûÿ¿øûÿ¿\032üÿ¿&üÿ¿0üÿ¿óýÿ¿\022þÿ¿,þÿ¿Aþÿ¿Xþÿ¿cþÿ¿\227þÿ¿¤þÿ¿¬þÿ¿¼þÿ¿Êþÿ¿Øþÿ¿éþÿ¿÷þÿ¿\002ÿÿ¿\rÿÿ¿@ÿÿ¿»ÿÿ¿")
> >    at lv_create_remove.c:42
> >42      inline int lv_create ( vg_t *vg, lv_t *lv, char *lv_name) {
> >(gdb) bt
> >#0  lv_create (vg=0x5, lv=0xbffffa14,
> >    lv_name=0xbffffa2c
> >"dûÿ¿\213ûÿ¿°ûÿ¿àûÿ¿øûÿ¿\032üÿ¿&üÿ¿0üÿ¿óýÿ¿\022þÿ¿,þÿ¿Aþÿ¿Xþÿ¿cþÿ¿\227þÿ¿¤þÿ¿¬þÿ¿¼þÿ¿Êþÿ¿Øþÿ¿éþÿ¿÷þÿ¿\002ÿÿ¿\rÿÿ¿@ÿÿ¿»ÿÿ¿")
> >    at lv_create_remove.c:42
> >#1  0x0804b03d in main (argc=5, argv=0xbffffa14) at lvcreate.c:791
> >#2  0x4006e177 in __libc_start_main (main=0x8049390 <main>, argc=5,
> >    ubp_av=0xbffffa14, init=0x8048e88 <_init>, fini=0x804b4d0 <_fini>,
> >    rtld_fini=0x4000e184 <_dl_fini>, stack_end=0xbffffa0c)
> >    at ../sysdeps/generic/libc-start.c:129
> >
> 
> _______________________________________________
> linux-lvm mailing list
> linux-lvm@sistina.com
> http://lists.sistina.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html

-- 
Sean Burford    x34135
ITS Systems Specialist
Adelaide University

  reply	other threads:[~2001-09-27 11:00 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2001-09-27 10:14 [linux-lvm] lvcreate core dump on large disks, vg name corrupted Sean Burford
2001-09-27 10:35 ` svetljo
2001-09-27 11:00   ` Sean Burford [this message]
2001-09-27 11:15     ` svetljo
2001-09-28 16:16     ` [linux-lvm] LVM deployment tips ? Jim Cromie

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=3BB306C3.29ECD29F@adelaide.edu.au \
    --to=sean.burford@adelaide.edu.au \
    --cc=linux-lvm@sistina.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.