Recent content by webbrowser

Guess what. Almost exactly TWO years later, I got the same error message (NFS Mount I/O Error in ESXi) while upgrading my NFS server from wheezy to jessie. Did some searching, this was the second hit in google. Wahey, my own post, so I tried it out... and nope. of course, I didn't modify...

@staticlag, I understand that Reallocated_Sector_Ct and Current_Pending_Sector are generally the attributes to look out for on mechanical drives. But how about SSDs? I'm a bit nervous with my Crucal M500 SSD because it already shows these after a few months. Are they of concern? 1...

Thanks guys. My 3 good drives say: 11 Calibration_Retry_Count 0x0032 252 252 000 Old_age Always - 0 200 Multi_Zone_Error_Rate 0x002a 100 100 000 Old_age Always - 19 11 Calibration_Retry_Count 0x0032 252 252 000 Old_age Always...

smartctl has this to say: # smartctl -a /dev/sdm smartctl 5.41 2011-06-09 r3365 [x86_64-linux-3.2.0-4-amd64] (local build) Copyright (C) 2002-11 by Bruce Allen, http://smartmontools.sourceforge.net === START OF INFORMATION SECTION === Model Family: SAMSUNG SpinPoint F4 EG (AFT) Device...

One of my hard disks is being excruciatingly slow, and I can't quite figure out why. Hope you can help. How slow is it? # hdparm -Tt /dev/sdm /dev/sdm: Timing cached reads: 15042 MB in 2.00 seconds = 7526.71 MB/sec Timing buffered disk reads: 26 MB in 3.26 seconds = 7.99 MB/sec...

Excellent, and many thanks to omniscence for catching my mistake. To answer what I'd set out to find at the start: All of the above is true.

No, didn't. Argh, that looks right!

You're right that the file data is cached, so the error isn't detected by mdsum until after i did zfs unmount and mount. In my later tests, I used zpool scrub instead, with no difference. It always seems to hang. I've tried to put my vdisk on / (ext4). Same results for me... stumped.

@omniscience. Oh! your results do look better than mine. What are you running? I'm on debian wheezy: # uname -a Linux storage 3.2.0-4-amd64 #1 SMP Debian 3.2.46-1 x86_64 GNU/Linux # dpkg -l zfs*|grep ^i ii zfs-dkms 0.6.1-1~whee amd64 Native ZFS filesystem kernel modu ii...

I re-ran my test, with these differences: - set failmode=continue - one instance of a one-byte data corruption, instead of 2 instances of 2048 byte corruption - used zpool scrub to detect the data corruption rather than just reading the files # dd if=/dev/zero of=/run/shm/vdisk bs=1024768...

I found this answer on Server Fault (http://serverfault.com/questions/381910/is-zfs-disk-rotting-protection-automatic-and-built-in), which basically says that metadata has more than one copy, and can self heal. Whereas file data does not unless we specify redundancy options. Right now: - I...

Thanks TCM. Failmode is currently set to wait. Would setting failmode=continue help? From my reading of the manpage, - writes would block. So I wouldn't be able to restore the files from backup - reads may be made to healthy devices. but I only have one device on this pool, and it seems...

oh. would it be different on solaris/bsd?

I'm wondering what happens if i have a single device zpool, and silent data corruption occurs. Does anyone know? What I'd expect is - data corruption is detected. - it can't be healed automatically since there is no redundancy What I'd ideally like to happen is - all files that are not...

Can't seem to delete my post.... Found my problem. Had left an entry in /etc/hosts.deny from before i had decided to just use iptables to firewall it off.