ZFS: third time through

Fascinating–I had just finished a scrub on June 6. Then a drive started dying. Now I’ve attempted to replace it, but an adjacent drive also had errors. I feel like I’m in a bit of a pickle. This snapshot had the error, so I deleted the snapshot and did a zfs clear tank, and the scrub process automatically restarted:

tank/VMs/l_4548-f30m64r@0000-installed:/lf541-f30-64-sda.img
  pool: tank
 state: DEGRADED
status: One or more devices is currently being resilvered.  The pool will
	continue to function, possibly in a degraded state.
action: Wait for the resilver to complete.
  scan: resilver in progress since Sat Jul  4 09:31:08 2020
	3.61T scanned out of 5.02T at 474M/s, 0h51m to go
	464G resilvered, 71.96% done
config:

	NAME                                                      STATE     READ WRITE CKSUM
	tank                                                      DEGRADED     0     0     0
	  mirror-0                                                DEGRADED     0     0     0
	    wwn-0x5000c500536ffa79                                ONLINE       0     0     0
	    replacing-1                                           DEGRADED     0     0     0
	      8130638507939855275                                 UNAVAIL      0     0     0  was /dev/disk/by-id/wwn-0x5000c5006e3d9d3f-part1
	      wwn-0x5000cca24cd7f690                              ONLINE       0     0     0
	  mirror-1                                                ONLINE       0     0     0
	    wwn-0x5000c5005226b37b                                ONLINE       0     0     0
	    wwn-0x5000c500522766b5                                ONLINE       0     0     0
	  mirror-3                                                ONLINE       0     0     0
	    wwn-0x5000cca223ca0714                                ONLINE       0     0     0
	    wwn-0x5000cca224cb6a3b                                ONLINE       0     0     0
	logs
	  mirror-2                                                ONLINE       0     0     0
	    nvme-SAMSUNG_MZVPW128HEGM-00000_S347NY0HB06043-part1  ONLINE       0     0     0
	    nvme-SAMSUNG_MZVPW128HEGM-00000_S347NY0HB06165-part1  ONLINE       0     0     0
	cache
	  nvme-eui.002538cb61020e02-part2                         ONLINE       0     0     0
	  nvme-eui.002538cb61020e7c-part2                         ONLINE       0     0     0

errors: Permanent errors have been detected in the following files:

        <0xb081>:<0x2>

So at the bottom it has a unnamed file reference and it is prolly going to stick at 71.x% for the next two hours.

ZFS Snapshot alias

Add this to your .bash_aliases for fun and profit:

function Snapshot () {
  local dst=""
  local atnam=""
  if [ -z "$1" ]; then
    dst=`df -l . | tail -1 |awk '{print $1}'`
  else
    if [[ $1 = *@* ]]; then
      atnam="${1##*@}"
      dst="${1%%@*}"
    fi
    dst=`df -l "$dst" | tail -1 |awk '{print $1}'`
  fi
  [ -z "$dst" ] && echo "wants file system name to snapshot" && return 1
  local NOW=`date +%Y%m%d-%H%M%S`
  [[ $dst = /* ]] && dst="${dst#/}"
  [[ $dst = */ ]] && dst="${dst%/}"
  [[ x$atnam = x ]] && atnam=$NOW
  sudo zfs snapshot "${dst}@${atnam}"
}

 

ZFS Rebuild Script

I’ve rebuilt my zfs modules often enough that I’ve written a script to do a clean build that should avoid old kernel modules and old libraries.

#!/bin/bash
sudo find /lib/modules -depth -type d -iname "spl" -exec rm -rf {} \;
sudo find /lib/modules -depth -type d -iname "zfs" -exec rm -rf {} \;
sudo find /usr/local/src/ -type d -a \( \
   -iname "spl-*" \
   -o -iname "zfs-*" \
   \) -exec rm -rf {} \;

sudo find /usr/local/lib/ -type f -a \( \
   -iname "libzfs*" \
   -o -iname "libzpool*" \
   -o -iname "libnvpair*" \
   \) -exec rm -f {} \;

cd spl
git reset --hard HEAD
git checkout master
git pull
git tag | tail -1 | xargs git checkout
./autogen.sh && ./configure && make -j13 && sudo make install
cd ../zfs
git reset --hard HEAD
git checkout master
git pull
git tag | tail -1 | xargs git checkout
./autogen.sh && ./configure && make -j13 && sudo make install

sudo update-initramfs -u
sudo update-grub2

Build OpenZFS on Ubuntu 16.04 from git

I have to import a zpool from one Ubuntu workstation with recent features to a new Ubuntu workstation. The new workstation being a fresh Ubuntu 16.04 Server install. It only has ubuntu-mate-desktop and build-essentials installed. Below is an aggregation of some of the apt install commands I preformed to get things going:

apt install dkms
apt install automake autoconf
apt install uuid-dev
sudo apt install libblkid-dev
sudo apt install -y libattr1-dev
apt install libnvpair1linux

This should get to to the point where you can do these commands:

$ git clone https://github.com/zfsonlinux/spl
$ ./configure && make -j13 && make install
$ git clone https://github.com/zfsonlinux/zfs
$ ./configure && make -j13 && make install
$ update-grub2

Notes on Updating Centos7 to 4.3.3

Upgrades! Sometimes they are a lot of homework.

Enabled Centos-Plus repos and elrepo for recent kernels. I figured out I want to install kernel-ml, kernel-ml-headers, kernel-ml-devel. That last one escaped me but is necessary because you need it when you do a dkms install.

So after updating that stuff, I was able to dkms install spl/0.6.5.3, dkms install zfs/0.6.5.3. I also made sure to modprobe spl zfs. Linked /usr/lib/systemd/system/zfs.target into /etc/systemd/system/sysinit.target.wants. Reboot. Make sure zfs pool returns.

To get vboxdrv all working I needed to make sure I installed the vboxdriver. That is /usr/sbin/rcvboxdrv. Easy to forget, that’s rather new. Then (as root) do a vboxmanage extpack install ~/Downloads/VirtualBox-extpack-5.0.6.extpack and that will keep your guest vms from blocking on missing USB-UHCI drivers.

Then go back and read my prev post on updating grub. I kept booting into kernel 3.10.x and wondering why grub2-install /dev/sda wasn’t doing it’s job. Ubuntu did the right amount of work with the update-grub2 script, I’ll say.

Ubuntu 15.10, ZFS 0.6.5.3…Fragile

Quick update on using a more recent version of zfs:

  • very glad I set up a user who’s home directory is not /home, because if zfs doesn’t finish a scan, no /home
  • attempting to install zfsnap, simplesnapshot, things that depend on zfsutils…all mess with kernel module
  • this was much easier back when there were less options and ppa:zfslinux-stable was available, didn’t have near this amount of difficulty
  • glad that zfs-auto-snapshot, zxfer are pretty easy to install with a Makefile
  • intrigued by zfSnap and simplesnapshot as backup tools but damned if I’m going to install them again after my .ko’s got all messed up.
  • Was hoping that 0.6.5.3 would have been promoted into 15.10 by now.

So, here’s hoping to getting a properly booting system :-)

Ubuntu 15.10 and ZFS

Screenshot-root@cholla:~

Some quick thots on doing this for my workstation:

  1. I have six 2TB drives in raid 10 zfs pool, and they would not import to 15.10 because 15.10 ships with (or tries to) zfs 0.6.4.2
  2. I decided on /boot, swap, /, mdadm partitions for OS install
  3. needed to do 15.10 server cmdline install for mdadm raid setup
  4. glad to not have attempted zfs-on-root for this distro
  5. setup three extra partitions on my two 120GB SSDS, using them for
    1. tank zil
    2. tank l2arc
    3. home pool (second pool named homer :-)
  6. Do not attempt to use PPA ubuntu/zfs-stable anymore, 15.10 will not accept it and it WILL mess with your zfsutils-linux recommended install.
  7. Somehow it did end up installing zfs-fuse. Somehow trying to install spl-dkms and zfs-dkms and uninstalling zfsutils-linux apt-get chose it. Why?
  8. I purged zfsutils, zfs/spl-dkms and did git clones on github/zfsonlinux/{spl,zfs}
  9. All of this required starting off with build-essential, autotools, automake, auto… and libuuid and … stuff. Not difficult to chase down.
  10. ./autoconfig.sh, ./configure && make -j10 && make install for spl and zfs
  11. updated /etc/rc.local to modprobe spl zfs, zpool import tank; zpool import homer; zfs mount tank ; zfs mount homer

I am able to reboot and import without pool version warnings.

Why did I move off 14.04.x? I really want to do video editing for kid videos and all the video packages for 14.04 are way ancient.

Also:

  1. get first server install working
  2. install lubuntu-desktop
  3. replace /etc/default/grub hidden -> false
  4. default/grub -> replace “splash quiet” with “nofb”
  5. once LXDE displays, then I do a “apt-get install mate-desktop-*” which seems to work just fine.
  6. Why? lubuntu-desktop flawlessly sets up Xorg dependencies and gives me a desktop the first time without messing around wondering why mate-desktop didn’t.

Merry Xmas!

Crazy Times with zxfer

I’ve started using zxfer that @AllanJude referred me to recently. It does a nice job. My main difficulty was how to get it to work efficiently over the 10Mbps that’s my effective DSL speed.

First, I made a copy of zxfer (zmxfer) that incorporates mbuffer. This is a crude hack, but helps me ensure that I’m getting around the mysterious hanging transmits I have previously seen sending zfs to zfs. Mbuffer seems to smooth this out well.

$LZFS send -i "$copyprev" "$copysrc" \| \
/usr/local/bin/mbuffer -q -s 128k -m 128M \
| /usr/local/bin/mbuffer -q -s 128k -m 128M \
| $RZFS receive $option_F "$copydest" \
|| { echo "Error when zfs send/receiving."; beep; exit 1; }

My off-site transfer script ssh’s to the primary backup server, queries a list of zfs filesystems to replicate and copies that back:

#~/bin/bash
CMDLIST=/tmp/zxfer_cmds.txt
XFPRE=/tmp/zxfer_batch_
SK=.ssh/backup_dsa
rm -f /tmp/zxfer_cmds*
if [ `ls /tmp/xfer-* 2>/dev/null | wc -l` -gt 0 ] ; then
   echo "Previous transfer in progress, bye."
   exit 1
fi
ssh -i $SK juno ./mk_fs_list.sh || \
   ( echo "Crap, didn't generate file-system list, bye."; exit 1 )
scp -i $SK juno:/tmp/vol_list /tmp || \
   ( echo "Crap, didn't copy file-system list, bye."; exit 1 )

We need to turn that list of filesystems into actual transfer commands. I create a file that full of the commands to execute later:

while read FS ; do
   [ -z "$FS" ] && continue;
   PFS=`dirname $FS`
   if [ "$PFS" == "." ] ; then 
      PFS=tank
   else
      PFS="tank/$PFS"
   fi
   echo "[ ! -f /tmp/stop-xfer ] && sudo zmxfer -dFPsv \
 -O \"-i .ssh/backup_dsa ctbu@juno sudo \" \
 -N tank/$FS $PFS"
done < /tmp/vol_list > $CMDLIST

You might think, “what a lot of sudo!” It’s good practice. I have dedicated a backup user to do this instead of root. I’ve configured the necessary sudoers file entries to make this work.

TIP: disable requiretty in sudoers [S.O.]

We want to increase the parallelism of these zfs transfers as much as possible. The time it takes to transfer zero-length snapshots in serial is prohibitive.

L=`wc -l < $CMDLIST`
Q=$[ $[ $L + 8 ] / 8 ]
split -l $Q $CMDLIST $XFPRE

Now we run these in screen, partly because ssh and sudo and mbuffer all tend to get a bit grouchy if they can’t agree on if the really need a tty or not…and mostly because I want to keep tabs on where any transfer hangups are. This keeps script output collated. First we test for and fire up a detached screen as necessary:

screen -ls xfer | fgrep -q '.xfer' || screen -dmS xfer
sleep 1

And then we fill the screen with some commands. (We need to have a .screenrc that defines eight screens.)

i=0
for x in $XFPRE* ; do
   echo "rm /tmp/xfer-$i" >> $x
   cmd="touch /tmp/xfer-$i"
   screen -S xfer -p$i -X stuff $"$cmd\n"
   screen -S xfer -p$i -X stuff $"time bash -x $x\n"
   i=$[ $i + 1 ]
done

Once this pxfer-lists.sh script of mine is run, you can connect to the screen using:

screen -S xfer -x

And watch the scripts do their stuff. (That stuff command is actually a true screen directive: stuff $crap into terminal $p.)

I’ve been able to get my transfer time down from 140 minutes to about 14 minutes. Also many of the backups I started transferring I figured out how to reduce in scope by stopping hourly snapshots on file systems that don’t require them.