CHANNEL(PAYMENT.GATEWAY.01) STATUS(RUNNING)
AMQ6125E wasn’t a wall. It was just a very confusing door.
She opened a second terminal. Checked the channel status: CHANNEL(PAYMENT.GATEWAY.01) STATUS(RETRYING) . Then the authentication records: SET CHLAUTH(PAYMENT.GATEWAY.01) TYPE(SSLPEERMAP) SSLPEER('CN=gateway-old,OU=payments') . Old certificate. The container cluster was using CN=gateway-new . But the queue manager had cached the SSL context after a partial renegotiation and—according to the FDC—tried to free a memory pointer it had already freed. amq6125e an internal ibm mq error has occurred
Lena didn’t call IBM support. She’d be on hold for an hour. Instead, she killed the channel process manually—not the channel, but the underlying amqrmppa process on the queue manager side.
Lena typed back: “Internal error. Fixed with forceful disagreement.” CHANNEL(PAYMENT
She closed her laptop, walked to the break room, and poured cold coffee into a mug. Outside, the city was still dark. Somewhere in the IBM MQ source code, line 2,417 of amqzfchk.c still had a flaw. But tonight, it didn’t matter.
Her phone buzzed. The on-call director: “Why is the payment retry queue frozen?” Checked the channel status: CHANNEL(PAYMENT
That was it. A double-free in the handshake logic. The queue manager had essentially stabbed itself in the back.