activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Paul Gale <paul.n.g...@gmail.com>
Subject Re: ActiveMQ 5.8.0: java.io.EOFException: Chunk stream does not exist, page: 19 is marked free
Date Wed, 10 Sep 2014 16:33:16 GMT
All of the following is assuming you're using Linux. I'm using RHEL 6.3 to
mount an NFSv3 based device using autofs.

I should have added that the issue for me was that I had specified the
wrong block size values for the rsize/wsize parameters in the autofs mount
configuration for the device I was mounting.

I was operating in the mistaken belief that the larger the value for these
parameters the better. Therefore I set their values to be 256K (in bytes).
Problems with the message store followed.

What I should have done, and ended up doing, was rather than guess the
value of the device's block size was to determine the device's _actual_
block size. You can either ask the administrator for the device what its
block size is or use the stat command.

If you really want to play it safe you could always use the default block
size for a device that supports NFSv3 which is quite conservative at 8192
bytes (I think - look it up). However, if the device you're mounting can
support larger block sizes then the stat command is how you would find that
out.

First, mount the device using a _very_ conservative block size value, say
1024 bytes. Second, run the stat command on the mount point to see what the
device's block size actually is. It might be the default 8192 or it could
be larger. Either way you'll know.

Here's an example. Say your local mount point is /NFS then the stat command
to use is:

stat -f /NFS

The output should look something like:

File: "/NFS"
ID: 0        Namelen: 255    Type: nfs
Block size: 32768      Fundamental block size: 32768
Blocks: Total: 330424288  Free: 178080429  Available: 178080429
Inodes: Total: 257949694  Free: 246974355

The output indicates the block size in bytes (32768) for the device. This
is the value that should be plugged into the rsize/wsize parameters for the
mount's definition.

I hope this helps.

Thanks,
Paul


Thanks,
Paul

On Wed, Sep 10, 2014 at 10:27 AM, Paul Gale <paul.n.gale@gmail.com> wrote:

> In my particular case I fixed it when I realized that I had the NFS mount
> settings for the mount where the KahaDB message store was located
> mis-configured. Since correcting the settings I've not had a single
> problem.
>
> Are you using NFS?
>
>
> Thanks,
> Paul
>
> On Tue, Sep 9, 2014 at 2:49 AM, khandelwalanuj <
> khandelwal.anuj90@gmail.com> wrote:
>
>> I am also seeing the same exception with ActiveMQv5.10. It comes
>> infrequent
>> and non-reproducible.
>>
>> I have already posted
>>
>> http://activemq.2283324.n4.nabble.com/ActiveMQ-exception-quot-Failed-to-browse-Topic-quot-td4683227.html#a4683305
>>
>>
>> ActiveMQGods can you please help us out here. ?
>>
>>
>>
>> Thanks,
>> Anuj
>>
>>
>>
>> --
>> View this message in context:
>> http://activemq.2283324.n4.nabble.com/ActiveMQ-5-8-0-java-io-EOFException-Chunk-stream-does-not-exist-page-19-is-marked-free-tp4672029p4685397.html
>> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>>
>
>

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