It seems that the cleanup on the journal is being called before the Journal was even started. That was maybe a situation that was caused by the earlier bug..
Let me know if you will be able to share the data or not.. I would definitely look at the data to see what caused it.
ive uploaded all the data files (its not that bad once ive tar'ed it - around 160mb)
here is the link
i hope it helps (i can also upload the print-data and print pages but i think you may get more out of those files)
I will get into it this Thursday morning
any chance you had time to check this out?
Im testing the new hornet and things are looking really good
I just tried to load your journal (I didn't even read its content for the record), and it started correctly.
there were a few page caches that couldn't be found (Instead of throwing the NPE), what was caused probably by the bug that it's fixed on this current release:
[^[* [main] 24-Jan 14:3:8,65 WARNING [PageSubscriptionImpl] Couldn't find page cache for page PagePositionImpl [pageNr=7825642, messageNr=-1, recordID=7678377343], removing it from the journal
* [main] 24-Jan 14:3:8,74 WARNING [PageSubscriptionImpl] Couldn't find page cache for page PagePositionImpl [pageNr=6056410, messageNr=4, recordID=7461713594], removing it from the journal
* [main] 24-Jan 14:3:8,75 WARNING [PageSubscriptionImpl] Couldn't find page cache for page PagePositionImpl [pageNr=6056410, messageNr=5, recordID=7461713637], removing it from the journal
* [main] 24-Jan 14:3:8,75 WARNING [PageSubscriptionImpl] Couldn't find page cache for page PagePositionImpl [pageNr=6056410, messageNr=8, recordID=7461713754], removing it from the journal
* [main] 24-Jan 14:3:8,79 WARNING [PageSubscriptionImpl] Couldn't find page cache for page PagePositionImpl [pageNr=8519740, messageNr=-1, recordID=7678377239], removing it from the journal
You should definitely get the latest release.
If you are still seeing issues with those already deleted.. it's probably because of something on the embedded. (either your class.. or our class that we delivered you). If you're still seeing those issues I will need some code from you starting this data and replicating that other issue.
Ok so ive got some bad news... this morning, the queues stopped delivering again. I dont have access to the data files right now, but ill analyze them later (and maybe send them to you if you're interested)
Ill try to start hornetq as a standalone server (and not embedded), but i dont think that is the problem - i am literally just starting the EmbeddedJMS class but maybe some other service needs to be configured. I have 1 AddressSettings, 2 Acceptors (one for the in-vm and the other a netty), a discovery and a broadcast services for cluster discovery, two connection configuration (again, an in-vm one and a netty one) and finally the queues and a topic.
But are you using the latest version on that server?
Yeah. i am using the tip from the 2.2.24.EAP.snapshot. Ill upload the jounal and paging files later so i can share with you.
It's just weird that I couldn't replicate the same scenario as you had ... So I'm wondering about the embedded.
You wouldn't be able to deliver messages even after a restart?
Yup. Even after a restart it would not resume the message delivery. Ill try to start hornet as a standalone later to see if there is any difference.
One thing that could be different are threading and some configuration?
So sorry cleber, it was my mistake. I was still running the previous version and i have not noticed until you said so. We are still testing with the snapshot but it seems to be working.
For what is worth, the previous version also locks in standalone mode and the new version works fine either standalone or embedded.
Ill get back to you regardless. Many thanks for all the help and patience
I will need logs and the data If something happens again. But I really believe you should be good now.
Ok just to put an end to this: the new version indeed solved all of our problems and we're quite happy. Thanks for the all the help and sorry for the inconvenience