Couple of questions: - How exactly are you attempting to delete the message? - What exactly are you doing when you remove the 0-byte large message file? Are you stopping the broker first? At what point is the 0-byte file recreated? - How do you know the reference still exists in the journal after you delete the message using the web console or CLI?
Justin On Thu, May 14, 2026 at 5:39 AM André António Santos Antunes Conrado <[email protected]> wrote: > > Hi, > > We have an ActiveMQ Artemis cluster with three nodes, previously running > version *2.42.0*. The cluster has generally been stable, but recently we > noticed an unusual issue. > > Inside .../artemis-broker/data/large-messages, we found several zero-byte > large message files, for example: > > -rw-r--r--. 1 artemis artemis 0 Mar 30 13:25 13017812601.msg > -rw-r--r--. 1 artemis artemis 0 Mar 30 13:25 13017812608.msg > -rw-r--r--. 1 artemis artemis 0 Mar 30 13:25 13017812613.msg > -rw-r--r--. 1 artemis artemis 0 Mar 30 13:25 13017812615.msg > -rw-r--r--. 1 artemis artemis 0 Mar 30 13:25 13017812622.msg > -rw-r--r--. 1 artemis artemis 0 Mar 30 13:25 13017817681.msg > > These messages appear to belong to internal inter-broker queues such as: > > $.artemis.internal.sf.artemis-cluster.9188e6b4-8238-11ef-8993-06b0c3061c67 > $.artemis.internal.sf.artemis-cluster.41e45e74-8238-11ef-91fe-023522aac679 > > For security reasons, we have since upgraded the cluster to version *2.53.0* > . > > At this point, recovering these messages is no longer important, although > we are not entirely sure whether they were successfully delivered > previously. > > The main issue is that we are unable to remove them. > > When deleting the messages through the web console, we receive the > confirmation: > > > "Messages deleted successfully [13017812601]" > > However, the message still remains afterward. We observe the same behavior > when using the CLI. > > We also attempted to delete the corresponding large message files manually, > but because references still exist in the journal, the files are > automatically recreated. > > What would be the safest and recommended procedure to permanently remove > these messages from the cluster and clean up the related journal references? > > Best regards, > > --- > André Antunes Conrado --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected] For further information, visit: https://activemq.apache.org/contact
