# GN3 Setup on tux01 - Non-Deterministic Behavior

## Tags

* assigned: bonfacem, zachs
* priority: high
* type: ops, guix
* keywords: gn3
* status: closed

## Description

Zach and Munyoki attempted to set up GN3 on tux01 following the provided channels file:

```
(list (channel
       (name 'gn-bioinformatics)
       (url "https://git.genenetwork.org/guix-bioinformatics")
       (branch "master")
       (commit
	"25cb9ad70155a08c088ff28bf54c06e12293105b"))
      (channel
       (name 'guix-forge)
       (url "https://git.systemreboot.net/guix-forge/")
       (branch "main")
       (introduction
	(make-channel-introduction
	 "0432e37b20dd678a02efee21adf0b9525a670310"
	 (openpgp-fingerprint
	  "7F73 0343 F2F0 9F3C 77BF  79D3 2E25 EE8B 6180 2BB3")))))

```

...  and execution steps:

```
source /var/guix/profiles/per-user/aruni/current-guix/etc/profile

guix pull -C channels.scm -p ~/.guix-extra-profiles/genenetwork

.  .guix-extra-profiles/genenetwork/etc/profile

guix install genenetwork3 genenetwork2 -p ~/.guix-extra-profiles/genenetwork

git clone https://github.com/genenetwork/genenetwork3.git

cd genenetwork3/

/export2/local/home/zas1024/gn3-zach/genenetwork3/run_dev.sh run --port=8090
```

Despite our efforts, attempting a CURL on GN3 using Zach's username consistently resulted in an empty response.

Even when Zach switched to my user (bonfacem) and repeated the steps with my ".guix-extra-profile", the CURL still produced an empty result.  We tried various troubleshooting steps, such as running the commands in a shell with no specific environment set up and using su to switch to bonfacem from Zach's userspace, but the issue persists.

## Resolution

The error caused was due to missing SSL certificates.  This was resolved by exporting the right SSL paths before starting the server.  The simple fix was adding:

```
export SSL_CERT_DIR="$GUIX_PROFILE/etc/ssl/certs"
export SSL_CERT_FILE="$GUIX_PROFILE/etc/ssl/certs/ca-certificates.crt"
export GIT_SSL_CAINFO="$SSL_CERT_FILE"
```

Also, debugging was difficult because of poor logging.  This is being tracked here:

=> issues/add-logging-to-rdf-endpoints.gmi

* closed