Is seaweedfs suitable for the back-end storage of mariadb and prometheus? #4241
Comments
|
how to reproduce? |
|
For example, mariadb： helm upgrade --install -n seaweedfs-test --create-namespace mariadb /opt/deploy/helm/mariadb -f /opt/deploy/helm/mariadb/values.yaml helm chart： values.yaml： |
|
Does it work with creating a simple file? |
|
Seems the number of volumes may be too low. |
|
it works |
|
the command you showed only used |
|
I can not reproduce by taking your values.xml screenshot and turn into a file. |
the screenshot is right |
|
|
I am also experiencing issues using seaweedfs as storage for prometheus. It looks like prometheus tries to do a poll which is not supported by seaweedfs. The following was run on ubuntu 23.10 with kernel Minimal reproduction steps# Start in whatever test directory you want
# Create backing directories
mkdir data mnt
# Start seaweedfs in the background
weed server -filer -dir=data &
# Mount seaweedfs in the background
weed mount -filer=localhost:8888 -dir=mnt -filer.path=/ &
# Create a directory for prometheus data inside the seaweedfs mount
mkdir mnt/prom
sudo chown 65534:65534 mnt/prom
# Try to start prometheus using seaweedfs for storage
docker run --rm -it -v "$(pwd)/mnt/prom:/prometheus" prom/prometheus:v2.48.1The final The prometheus logs: |
|
I can also confirm that trying to use seaweedfs as the storage backend for a mariadb database also fails in a similar way (though without the POLL error message). Using mysql itself does not seem to have the same issue. docker run --rm -e 'MARIADB_ROOT_PASSWORD=test' -v $(pwd)/mnt/mariadb:/var/lib/mysql mariadb |
|
I think this is related: hanwen/go-fuse#501 |
|
works on Mac: |
|
Hi Chris, is this with seaweedfs 3.55? And I assume this is APFS; i wonder if it has something to do with the underlying filesystem, or even the kernel. My original post was using an ext4 filesystem. I went ahead and made an XFS filesystem and ran the same test, and it also failed. Either way, I saw your post in go-fuse and I think the convo is better continued there. Thanks! |
|
Hi, go-fuse contributor here. This seems to be where things go wrong:
I don't see how this relates to POLL. Looks more like a READ is failing. But in both cases, a FUSE debug log would help. |
|
It's very possible that the POLL is a red herring... What's the best way for me to get the logs you're requesting? EDIT: Sorry, meant to continue this conversation in the other thread, feel free to respond there instead. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Describe the bug
use seaweedfs as the storage of kubernetes，mariadb and promethues failed to start.
if use nfs, mariadb and promethues started successfully.
System Setup
List the command line to start "weed master", "weed volume", "weed filer", "weed s3", "weed mount".
3 master、4 volume
nohup /usr/local/bin/weed master -ip=10.88.206.77 -port=9333 -defaultReplication="001" -mdir=/data1/weedfs/master -peers=10.88.206.78:9333,10.88.206.79:9333 > /var/log/weedfs/master/master.log 2>&1&
nohup /usr/local/bin/weed volume -port=10002 -dir=/data0/seaweedfs/volume/volume01 -max=7 -fileSizeLimitMB=200 -mserver=10.88.206.77:9333,10.88.206.78:9333,10.88.206.79:9333 -dataCenter=dc1 -rack=rack1 > /var/log/seaweedfs/volume2.log 2>&1&
nohup /usr/local/bin/weed filer -port=8888 -master=10.88.206.77:9333,10.88.206.78:9333,10.88.206.79:9333 -ip=10.88.206.77 > /var/log/weedfs/filter/filter.log 2>&1&
nohup /usr/local/bin/weed s3 -config=/etc/seaweedfs/s3-config.json -filer=10.88.206.77:8888 -ip.bind=10.88.206.77 -port=8333 > /var/log/weedfs/s3/s3.log 2>&1&
OS version：ubuntu20.04
weed version：3.42
filer.toml：
[leveldb2]
enabled = true
dir = "/data1/weedfs/filer/data"
Screenshots
this warn does not affect mariadb startup.
The text was updated successfully, but these errors were encountered: