Nx2l's Home Lab blog

Community Blog
,
41

So… im still messing with this crap…

Found the solution… try to patch now.

https://github.com/zfsonlinux/zfs/issues/8318

2 Likes
42

I default to rhel world as well, but if ZoL is in the mix, I go with Ubuntu for the native support.

I understand though. You gotta keep up appearances.

2 Likes
43

There must be something missing.
Not having any luck with those steps yet…

ll /lib/modules/4.20.10-200.fc29.x86_64/
total 16940
-rw-r–r--. 1 root root 258 Feb 15 13:57 bls.conf
lrwxrwxrwx. 1 root root 40 Feb 15 13:57 build -> /usr/src/kernels/4.20.10-200.fc29.x86_64
-rw-r–r--. 1 root root 201371 Feb 15 13:56 config
drwxr-xr-x. 2 root root 42 Feb 22 22:28 extra
drwxr-xr-x. 14 root root 157 Feb 22 21:13 kernel
-rw-r–r--. 1 root root 1074886 Feb 22 22:28 modules.alias
-rw-r–r--. 1 root root 1048515 Feb 22 22:28 modules.alias.bin
-rw-r–r--. 1 root root 1692 Feb 15 13:57 modules.block
-rw-r–r--. 1 root root 8296 Feb 15 13:57 modules.builtin
-rw-r–r--. 1 root root 10721 Feb 22 22:28 modules.builtin.bin
-rw-r–r--. 1 root root 394505 Feb 22 22:28 modules.dep
-rw-r–r--. 1 root root 544122 Feb 22 22:28 modules.dep.bin
-rw-r–r--. 1 root root 363 Feb 22 22:28 modules.devname
-rw-r–r--. 1 root root 153 Feb 15 13:57 modules.drm
-rw-r–r--. 1 root root 69 Feb 15 13:57 modules.modesetting
-rw-r–r--. 1 root root 2667 Feb 15 13:57 modules.networking
-rw-r–r--. 1 root root 136744 Feb 15 13:57 modules.order
-rw-r–r--. 1 root root 562 Feb 22 22:28 modules.softdep
-rw-r–r--. 1 root root 452719 Feb 22 22:28 modules.symbols
-rw-r–r--. 1 root root 553036 Feb 22 22:28 modules.symbols.bin
lrwxrwxrwx. 1 root root 5 Feb 15 13:57 source -> build
-rw-------. 1 root root 4111527 Feb 15 13:56 System.map
drwxr-xr-x. 2 root root 6 Feb 15 13:54 updates
drwxr-xr-x. 2 root root 40 Feb 22 21:13 vdso
-rwxr-xr-x. 1 root root 8753352 Feb 15 13:57 vmlinuz

Its not showing up… I guess i have to wait until the next zfs release… or go back a kernel…

im just going to stop tonight… for real this time.

Edit 1

Actually got it…
I imported my pool and now a disk is not showing up…

EDIT 2 :
The steps are:

  • Make sure that you are running the kernel that you want to build the module for:
    uname -srm Linux 4.20.7-100.fc28.x86_64 x86_64

  • Reinstall zfs and spl packages to have a clean environment:
    dnf reinstall zfs-dkms spl-dkms zfs

  • Download ubuntu’s zfs-linux tarball:
    wget https://mirrors.edge.kernel.org/ubuntu/pool/main/z/zfs-linux/zfs-linux_0.7.12-1ubuntu5.debian.tar.xz

  • Extract the archive
    tar xjf zfs-linux_0.7.12-1ubuntu5.debian.tar.xz

  • Enter the patches dir:
    cd debian/patches

  • Apply the patch to the zfs source:
    sudo patch -p1 /var/lib/dkms/zfs/0.7.12/source/include/zpios-ctl.h <3204-Add-4.20-timespec-compat-fix.patch

  • Remove the modules that dnf built upon installation of the package:
    sudo dkms remove spl/0.7.12 --all
    sudo dkms remove zfs/0.7.12 --all

  • Build the modules anew from the patched source code:
    sudo dkms --force install spl/0.7.12
    sudo dkms --force install zfs/0.7.12

  • Load the module: modprobe zfs
    Now you should be able to restart the zfs-services and mount your pool(s), or just reboot your machine to make the changes to take effect.

1 Like
44

I’m in the same boat. Was working on a jbod. Plugged it back in and half the drives are degraded or unavailable. Think this guy might be on the way out.

1 Like
45

I think we’re all clear now…
Just have to restore a few config files

image

.

image
image.png847×30 5.24 KB

.
.

I got the fibre targets enabled again to my compute server.

image

1 Like
46

Are you still up?

47

I woke up a couple hours ago

48

I’m still up :frowning: I also had a ZFS fiasco involving a resilver and subsequent scrub. SAS cables are so sensitive.

49

Sorry to hear.

what brand?

1 Like
50

The ones inside Supermicro JBODs. I’m not exactly sure what happened. Part of my fan mod was pinching or pulling or interfering with the connection in a way where the communication with the HBA was garbled.

It still saw about 2/3rds of my disks in varying states and tried to start resilvering over the bad connection. Fixed it and then it repaired itself and gave me s cryptic warning that “applications might be affected”.

51

Is everything ok now?

52

As far as I could tell everything was fine. There was a backup in the last 24 hours so not a shit-your-pants situation, I just needed to get the box back up.

1 Like
53

And then?

54

Alright team, I got another situation. I have some docker containers up and running, we’ll use keycloak at port 8082. When go to http://<ip>:8020, everything is fine and dandy.

Whomst-ever, I tried to add an nginx reverse proxy and it’s acting like it can’t find the host? My config basically looks like:

location / {
    proxy_set_header Host $host;
    proxy_set_header X-Real-IP $remote_addr;
    proxy_pass http://<ip>:8020;
}

I don’t remember the exact errors (I can get those tonight if it will help) but is there anything special about proxies in docker?

2 Likes
55

Wish I could help… but I’ve never messed with proxies in docker

1 Like
56

Im pretty interested in replacing my home lab setup an Epyc 7371 system… but other than a few articles…

I havent found anywhere that has it in stock.

1 Like
57

Hey guys, I thought you might like my Home(lab)-made Container Orchestration Platform that I made because (in part) of this blog. Lemme know what you think!

3 Likes
58

Considering selling my current storage setup (6 core i5 . 8th gen with 64GB memory) for this:

image
image.png642×568 77.2 KB

Of course reusing my SAS HBA, FC HBAs, PSU, and case.

Thoughts??

  • yes
  • no
  • yes, but…
  • no, but…
  • other …

0 voters

2 Likes
59

What will this allow you to do that you can’t do now?

2 Likes
60

maybe…but not really. My next step probably should be more harddrives…

damn you… asking the questions that make me reflect on my idea.

… might could if I also sold my compute system and integrated everything into one machine. But I dont think I want to do that.