Solved Haraka mailserver crashing after sending or receiving email

[ccfu]

I have noticed that the mailserver is crashing and immediately restarting regularly after successfully sending or receiving email. Anyone any idea why this is happening and / or what I can do to rectify it? This is what I get in the logs:

Feb 28 14:53:44 [INFO] [-] [core] [outbound] Socket [outbound::2424:127.0.0.1:undefined:300] in pool got FIN
Feb 28 14:53:44 [CRIT] [-] [core] Error: Resource not currently part of this pool
Feb 28 14:53:44 [CRIT] [-] [core] at Pool.destroy (/app/code/haraka/node_modules/generic-pool/lib/Pool.js:548:9)
Feb 28 14:53:44 [CRIT] [-] [core] at sockend (/app/code/haraka/outbound/client_pool.js:172:37)
Feb 28 14:53:44 [CRIT] [-] [core] at pluggableStream.<anonymous> (/app/code/haraka/outbound/client_pool.js:164:9)
Feb 28 14:53:44 [CRIT] [-] [core] at Object.onceWrapper (node:events:509:28)
Feb 28 14:53:44 [CRIT] [-] [core] at pluggableStream.emit (node:events:390:28)
Feb 28 14:53:44 [CRIT] [-] [core] at TLSSocket.<anonymous> (/app/code/haraka/tls_socket.js:79:18)
Feb 28 14:53:44 [CRIT] [-] [core] at TLSSocket.emit (node:events:402:35)
Feb 28 14:53:44 [CRIT] [-] [core] at endReadableNT (node:internal/streams/readable:1343:12)
Feb 28 14:53:44 [CRIT] [-] [core] at processTicksAndRejections (node:internal/process/task_queues:83:21)
Feb 28 14:53:44 [NOTICE] [-] [core] Shutting down
Feb 28 14:53:44 2023-02-28 13:53:44,398 INFO exited: haraka (exit status 1; not expected)
Feb 28 14:53:45 2023-02-28 13:53:45,407 INFO spawned: 'haraka' with pid 215445
Feb 28 14:53:45 loaded TLD files:
Feb 28 14:53:45 1=1487
Feb 28 14:53:45 2=8571
Feb 28 14:53:45 3=2448
Feb 28 14:53:45 loaded 9459 Public Suffixes
Feb 28 14:53:46 loglevel: INFO
Feb 28 14:53:46 log format: DEFAULT
Feb 28 14:53:46 Starting up Haraka version 2.8.28
Feb 28 14:53:46 [INFO] [-] [core] Loading plugins
Feb 28 14:53:46 [INFO] [-] [core] Loading plugin: delay_deny
Feb 28 14:53:46 [INFO] [-] [core] Loading plugin: dnsbl
Feb 28 14:53:46 [INFO] [-] [core] Loading plugin: helo.checks
Feb 28 14:53:46 [INFO] [-] [core] Loading plugin: headers
Feb 28 14:53:46 [INFO] [-] [core] Loading plugin: tls
Feb 28 14:53:46 [INFO] [-] [core] loading tls.ini
Feb 28 14:53:46 [INFO] [-] [core] Loading plugin: spf
Feb 28 14:53:46 [INFO] [-] [core] Loading plugin: cloudron
Feb 28 14:53:46 [INFO] [-] [core] Loading plugin: rcpt_to.in_host_list
Feb 28 14:53:46 [INFO] [-] [core] Loading plugin: dkim_sign
Feb 28 14:53:46 [INFO] [-] [core] Loading plugin: spamassassin
Feb 28 14:53:46 [INFO] [-] [core] Loading plugin: queue/smtp_forward
Feb 28 14:53:46 [INFO] [-] [core] Loading plugin: limit
Feb 28 14:53:46 [INFO] [-] [core] getting SocketOpts for SMTPS server
Feb 28 14:53:46 [NOTICE] [-] [core] Listening on [::0]:2525
Feb 28 14:53:46 [INFO] [-] [core] Creating TLS server on [::0]:2465
Feb 28 14:53:46 [NOTICE] [-] [core] Listening on [::0]:2465
Feb 28 14:53:46 [NOTICE] [-] [core] Listening on [::0]:2587
Feb 28 14:53:46 [INFO] [-] [cloudron] Initializing queue server on port 6000
Feb 28 14:53:46 [INFO] [-] [limit] connected to redis://127.0.0.1:6379/4
Feb 28 14:53:46 [INFO] [-] [core] [outbound] Loading outbound queue from /app/data/haraka-queue
Feb 28 14:53:46 [INFO] [-] [core] [outbound] Loading the queue...
Feb 28 14:53:46 [INFO] [-] [core] [outbound] [pid: undefined] 0 files in my delivery queue
Feb 28 14:53:46 [INFO] [-] [core] [outbound] [pid: undefined] 0 files in my load queue
Feb 28 14:53:46 [INFO] [-] [core] [outbound] [pid: undefined] 0 files in my temp fail queue
Feb 28 14:53:47 2023-02-28 13:53:47,261 INFO success: haraka entered RUNNING state, process has stayed up for > than 1 seconds (startsecs)

[Kubernetes]

I have also such crashes in my logs, but not after every e-mail sent or received, but sometimes.

[nebulon]

This might be related to https://github.com/haraka/Haraka/issues/3118

[ccfu]

@nebulon

Looks like it might be related, yes. Are there any plans to update the mailserver version in Cloudron soon? I don't think it is having a negative effect on mail operations but it is annoying. The error seems to be linked to pooling, which has been changed in version 3.

[girish]

The current stable version of Cloudron uses Haraka 83a952b009cda .

We already updated Haraka to 3.0.1 which maybe fixes this issue. The pooling code is now gone - https://github.com/haraka/Haraka/pull/3115 .

@ccfu @Kubernetes maybe we can try to update the mail container in your servers to see if things work better now? Otherwise, we have to debug this further. Can you write to us at support@cloudron.io ?

[girish]

Following up from support: server was patched manually to latest version of Haraka.

[opensourced]

@girish I also see this error on one of my cloudrons.. Will your patch soon be released in an update?

[girish]

@opensourced yes, should be out next week along with 7.4