Hi Guys,
I have a server running CentOS 4.2 with a 2.6.9-22EL-smp kernel.
The system contains a 2.3 TB RAID 5 array. There is a 20GB root partition with the remainder of the drive configured one large LVM volume. (/u0)
I have installed Veritas Netbackup on the server and am running a few tests. Small backups and restores work just fine.
When I run a larger backup (500MB) and store the backup file on /u0/vrts the files within the backup cannot be restored. I thought at first this was a Veritas problem but have also discovered...
I cannot do any of the following to the backup file: Use cat, dd or cp to copy the file. The first part of the file aproximately 24MB will copy and then I get the error message: Input/output error
Checking dmesg shows: dm-0: rw=0, want=31933816664, limit=4838391808 attempt to access beyond end of device dm-0: rw=0, want=32472792952, limit=4838391808 attempt to access beyond end of device dm-0: rw=0, want=33011769240, limit=4838391808 attempt to access beyond end of device dm-0: rw=0, want=33550745528, limit=4838391808 attempt to access beyond end of device dm-0: rw=0, want=34089721816, limit=4838391808
How do I access the files in question?
Any suggestions or enlightment would be greatly apreciated.
Shawn
A few relevant comments (inline below)
On Saturday 30 September 2006 07:10, Shawn Everett wrote:
Hi Guys,
I have a server running CentOS 4.2 with a 2.6.9-22EL-smp kernel.
Why not run a fully updated centos-4?
The system contains a 2.3 TB RAID 5 array. There is a 20GB root partition with the remainder of the drive configured one large LVM volume. (/u0) ... I cannot do any of the following to the backup file: Use cat, dd or cp to copy the file. The first part of the file aproximately 24MB will copy and then I get the error message: Input/output error
Booting from it implies that you have a "normal" partition table on the 2.3T device, this does not work (dos style MBRs only works with <2T). Most people boot from a small and independent device and uses gpt (see parted) on the big device or no table at all (lvm directly on the device).
Good luck, Peter
Checking dmesg shows: dm-0: rw=0, want=31933816664, limit=4838391808 attempt to access beyond end of device dm-0: rw=0, want=32472792952, limit=4838391808 attempt to access beyond end of device dm-0: rw=0, want=33011769240, limit=4838391808 attempt to access beyond end of device dm-0: rw=0, want=33550745528, limit=4838391808 attempt to access beyond end of device dm-0: rw=0, want=34089721816, limit=4838391808
Why not run a fully updated centos-4?
The system contains a 2.3 TB RAID 5 array. There is a 20GB root partition with the remainder of the drive configured one large LVM volume. (/u0) ... I cannot do any of the following to the backup file: Use cat, dd or cp to copy the file. The first part of the file aproximately 24MB will copy and then I get the error message: Input/output error
Booting from it implies that you have a "normal" partition table on the 2.3T device, this does not work (dos style MBRs only works with <2T). Most people boot from a small and independent device and uses gpt (see parted) on the big device or no table at all (lvm directly on the device).
Hi Peter,
I took your advice. I upgraded to CentOS 4.4 and then further upgraded the kernel to the official 2.6.18 kernel.
I installed a basic IDE drive and did the OS load on to that drive. The remaining disks were resetup using hardware RAID 5 and LVM was setup directly on the device (no partitions).
Also to avoid any issues with large file systems and EXT3 I setup and formatted the LVM with JFS. It's worth noting that a mkfs.jfs on 2.3TB took about 10 seconds! For some odd reason the CentOS kernel doesn't support JFS so a recompile was in order. Downloading and installing jfsutils was trivial.
Try as I might, I have not yet been able to reproduce the problems that caused me stress. I can create multi GB files without a problem. I can move files into and out of the partition without an error. I'm taking this as a good sign. I'm also not seeing any disk access errors in dmesg!
Shawn
On Sat, Sep 30, 2006 at 06:37:18PM -0700, Shawn Everett wrote:
I took your advice. I upgraded to CentOS 4.4 and then further upgraded the kernel to the official 2.6.18 kernel.
Try as I might, I have not yet been able to reproduce the problems that caused me stress. I can create multi GB files without a problem. I can move files into and out of the partition without an error. I'm taking this as a good sign. I'm also not seeing any disk access errors in dmesg!
I built an early RHEL 4 (maybe U1, maybe the original release) with a ~12TB raid system. Things were very, very sad with the rhel kernel, but updating to the then-current kernel made things much, much better.
I also, if i recall correctly, needed to get the most recent XFS tools, but I had to do that anyway since redhat doesn't support XFS.
danno -- Dan Pritts, System Administrator Internet2 office: +1-734-352-4953 | mobile: +1-734-834-7224
-
Dan Pritts -
Peter Kjellström -
Shawn Everett