StorJ "crashes" on RBPI 4 after some time

Node Operators
,
1

Hi,
i am running 2 storagenodes succesfull via Docker. Now, I started one on a RBPI 4 with 4GB of RAM and a external 6TB HDD.
I am running following config:

docker run -d --restart unless-stopped --stop-timeout 300 \
    -p 28967:28967/tcp \
    -p 28967:28967/udp \
    -p 0.0.0.0:14002:14002 \
    -e WALLET="PLACEHOLDER" \
    -e EMAIL="PLACEHOLDER" \
    -e ADDRESS="PLACEHOLDER:28967" \
    -e STORAGE="5TB" \
    --log-opt max-size=50m \
    --log-opt max-file=10 \
    --user $(id -u):$(id -g) \
    --mount type=bind,source="/storj/StorJ/Identiy/Identity/storagenode",destination=/app/identity \
    --mount type=bind,source="/storj/StorJ/Data",destination=/app/config \
	--mount type=bind,source="/storjdatabase",destination=/app/dbs \
    --name storagenode storjlabs/storagenode:latest

After some time, the node crashed. The CPU-Usage spikes to nearly 100% and the log freezes.
I also tried to set a Memory Limit to like 800mb but that didn’t helped. I also moved the Databse Files from the external HDD to a external SSD which is connected via USB3.

What causes this kind of crash?

I am running Ubuntu Server 22.04 ARM64

Thank You!

2

Try this one:

docker run -d --restart unless-stopped --stop-timeout 300 \
…
    -e STORAGE="9.5TB" \
    --memory=1000m \
    --log-opt max-size=1g \
    --log-opt max-file=5 \
    --mount …
    --name sn1 storjlabs/storagenode:latest

I store my log externally - pls add your path, if you want to do so.

I’ve put memory param to 1000m.

There’s an RPi3 dedicated setup guide, where there is a workaround for memory issues:

The tip begins on this website with: „ To add cgroup memory support (to prevent an OOM hang of your Pi 3 B/B+): …“

Also secure, that you use a CMR drive (!).

Anything the log gives you in terms of error messages?

1 Like
3

I think your maybe confusing cpu usage with IO usage. You said your running 2 nodes but didnt mention that its on 2 hard drives or on 2 different systems Are both nodes running on the same hard drive?

2 Likes
4

Thanks for the suggestions. I will try them.

The logs don‘t show anything abnormal when the node crashes…

5

I have two nodes running as a docker on dedicated X86 Hardware successfully, which is more poweful and doesn‘t produce any problems since 2 years…
The setup is one node per drive. I even moved the Database on all of them (including the PI) to a SSD to minimize IO issues.

But my latest 3. Node on the PI seems to have performance issues

6

I think you should check in dmesg for storage errors if you haven’t already.
Also try moving the disk to the slower usb2 port to get a comparison

3 Likes