activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Tim Bain <>
Subject Re: failed to start ActiveMQ
Date Mon, 12 Mar 2018 01:58:58 GMT
On Sat, Mar 10, 2018 at 12:50 AM, norinos <> wrote:

> Hi, Tim !
> Yes. I'm using 5.13.1.
> I will try debugging and check the location size.
> In this case, my KahaDB size is very large.
> It seems that the message in which the offlineDurableSubscriberTimeout
> occurred

Uh, sorry, context please? This was not mentioned in your initial post.

> has not been deleted,
> so I restart ActiveMQ then this trouble is occrred.
> Do you think that the message is not deleted also because KahaDB location
> size is broken?

I don't really know the KahaDB code well enough to answer that question. I
certainly can't rule it out, though it's quite possible that a copy of the
message that doesn't have the problem might be in the memory store and this
problem might occur only when loading persistent messages from disk on a
broker restart. Either is possible, but I'm not sure it really matters. The
ultimate problem is that either the message was invalid and we didn't
detect it on receipt, or it got corrupted sometime after that; either way,
if we fix that root cause, the inability to delete the message (if that's
indeed what's going on) won't matter. But first let's have you confirm that
an invalid message size is in fact what's going on, via the steps I
provided earlier, and then we can figure out where to go from there.

Also, were there any filesystem/storage issues with the disk contents of
the KahaDB files around the time this message would have been received? If
there's a meaningful possibility that the file was simply corrupted due to
something outside the broker's control, let's know that up front.


  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message