more I/O errors in Proxmox2

[tincboy]

I'm using Proxmox2 on two of my locations, one of them is operating fine, but in the other location, 3 of 11 servers are showing bad sector errors in syslog
In the location with bad sectors I have to use a pre-installed Debian 6, therefore the default file system is Ext4 .
this is a little strange to me and I wonder if this has something to do with the file system , kernel or proxmox .
hard disk :seagate
model :ST3000DM001-9YN166
partition :/dev/sda3
file system :EXT4

I'm looking forward to any suggestion about the origin of this problem and how to solve and prevent it.
Best Regards

 

[dietmar]

Bad sectors sound like disk failure, not software bug.

 

[tincboy]

I don't think it is a software bug, But I guess some mistaken configuration or incompatible config makes servers in the specific location fail, because all of them are using almost the same hardware configuration
Is ext4 compatible with Proxmox needs? I never had this issue in last 3 years with Proxmox nodes.

 

[dietmar]

Is ext4 compatible with Proxmox needs?

Yes. But again, that sound like a hardware/disk problem, not related to software.

 

[tincboy]

Thanks, Yes I agree the disk is now going to die and it's now a hardware issue.
But I need to know why the same kind of hardware was operating fine for more than a year, But just 3 month after I've start using Proxmox 2 many of my servers are experiencing bad sectors.

 

[dietmar]

But just 3 month after I've start using Proxmox 2 many of my servers are experiencing bad sectors.

Well, disks fails (unfortunately).

 

[Stewge]

Thanks, Yes I agree the disk is now going to die and it's now a hardware issue.
But I need to know why the same kind of hardware was operating fine for more than a year, But just 3 month after I've start using Proxmox 2 many of my servers are experiencing bad sectors.

Unfortunately, bad sectors are a physical degradation of hard drives. Nothing in software-land will prevent that, only things like SMART and RAID controllers which will silently map around these errors delay the inevitable. Bad sectors also seem to build exponentially. Once you get a few bad sectors, the problem gets exponentially worse until you replace it. This is why drives seem to fail "suddenly" with bad sectors even when you get warnings.

That being said, in the past 3-4 months I've had 1 SAN drive failure with bad sectors (hitachi ultrastar 2TB), my own personal backup drive (WD Green 2TB), a client server (WD yellow 1TB), a in-house VM server running Proxmox (WD Yellow 500GB) and a couple of client on-site servers on top of that. Out of all of these, the only BAD failure was my own backup drive since it wasn't RAID and just a USB external (so no SMART warnings or anything, just struggled copying files one day). Given that it was primarily a storage for various isos/applications it's nothing that can't be downloaded again.

My Theory:
It would seem, that prior to the hard-drive manufacturing crash, quality control must've just dropped immensely as demand was booming. Drives were rushed out incredibly fast (when 2TB was ~ $110-120). My suggestion, keep an eye on any SMART warnings and be prepared with backups and spare drives.