Troubleshooting

Troubleshooting

Start with the logs:

sudo journalctl -u hashtasker-server -f
sudo journalctl -u hashtasker-worker -f

Server

Symptom	Cause / Fix
config file 'config.yaml' not found	The server reads config from its working directory. Check WorkingDirectory in the unit file.
server.session_key must be set...	Set a real key: openssl rand -hex 32
database.password must be set...	Set any non-example value, even though SQLite is used
admin.password must be set...	Set a password other than changeme
Hash mode dropdown empty	hashmodes_path is wrong or hashmodes.txt is missing
Wordlist or ruleset dropdown empty	wordlists_path / rulesets_path wrong or unreadable
Changelog page empty or stale	Copy changelog.md into the working directory
UI looks old after an upgrade	Hard refresh (Ctrl+Shift+R)
Failed to migrate database	Restore your hashtasker.db backup and report the log output

Login

Symptom	Cause / Fix
"Invalid credentials" for LDAP users	The server log has the real reason. See Authentication and Users.
LDAP user is not admin	admin_group must match the full group DN in memberOf exactly. Nested groups do not count.
Logged out unexpectedly	Sessions last 7 days. Changing session_key invalidates all sessions.

Workers

Symptom	Cause / Fix
Worker never appears	Check server_url, firewall, and the worker log for connection errors
Worker flaps online/offline	checkin_interval too close to 60s, or the network is dropping requests
Two workers show as one	Duplicate hostname; make each unique
Worker shows queued	hashcat reported the host is overloaded (exit code 252). It retries automatically when a running job finishes.
No such file for a rule	The rule file name exists on the server but not in the worker's rulesets_path
Worker did not get part of a job	Only workers online at job creation get a chunk

Jobs

Symptom	Cause / Fix
"Please select a valid ... from the dropdown"	Typed text is not accepted; click an item in the dropdown
Copied job still used an old ruleset	Fixed in v1.3.2. Upgrade the server.
Job stuck in queued (NFS mode)	Line counting a very large wordlist is still running. Check the job's status message and the server log.
Job failed with an NFS message	A worker's mount dropped mid-run. Fix the mount and resubmit.
Cracked count lower than expected with potfile enabled	Expected: cracks are reported from each worker's potfile plus new ones. Disable the potfile to force a fresh crack.
"4 of 3 cracked" style counts	Fixed in v1.1.4. Upgrade.

NFS Mode

Symptom	Cause / Fix
"Cannot create job. NFS mount missing on ..."	The named host's wordlists_path is not readable. Remount, or take the worker offline.
Worker NFS shows "wordlists_path not configured"	Add wordlists_path to that worker's worker.json and restart it
Server NFS card red	The server's wordlists_path is not mounted or not readable