db-derby-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Myrna van Lunteren <m.v.lunte...@gmail.com>
Subject Re: problems after "could not listen on port xxx on host" / corrupting data base?
Date Thu, 24 Feb 2011 08:09:23 GMT
On Wed, Feb 23, 2011 at 2:33 PM, Morten Mikkelsen <mortenbreum@gmail.com> wrote:
> Hi,
> I'm really no expert, but I would say there is not enough information here..
> I know of no situations where you would get the 'address already in use'
> exception if the port is free.
> Perhaps the script used to start the network server occupies the port before
> starting the server, making the actual start fail?
> You should be able to start the server before worrying about the next
> problems, so focus on that first off.
> /Morten
> On Wed, Feb 23, 2011 at 10:50 PM, Thomas Hill <Thomas.K.Hill@t-online.de>
> wrote:
>> Hi Morten,
>> thanks for your reply "on the address already in user issue". The Network
>> Server
>> is hosted on the internet and my provider tells me that nothing else is
>> using
>> the port and their tests would indicate that - the normal (Linux)
>> start/stop
>> scripts coming with Derby are being used - when the Server is stopped the
>> port
>> is indeed properly released. They are assuming that if the port in a
>> (fresh)
>> start is not properly released the Derby Server might not have properly
>> been
>> shutdown / might have crashed - for which I don't have indications.
>> Any toughts on whether the server is started after the address already in
>> use
>> has been encountered? There a no further entries in the log that would say
>> 'server is ready to accept connections' - so I guess the start failed.
>> And any thougths on the 'another instance ...' issue - which prevents me
>> to boot
>> the database? might the database have been corrupted??
>> Tx
> --
> /Morten
I have seen this type of error pop up in rare situations - including
some bugs in networkserver (e.g. DERBY-4053, DERBY-4304), but also due
to jvm bugs, and from looking at DERBY-3108, it can be the result of a
normal TCP/IP limitation if you try to open and release connections
really fast. As Morten said, it's a classic network exception.

The fact that the database is not starting with ij is to *protect* it
from getting corrupted.

So...some questions I can think of:
What version of derby are you using?
What has happened between now and when you last could use the system -
did the system crash? Did someone ctrl-c or kill the network server
process? Did any other piece of software get installed on this system
that could be using the same port?
Can you start networkserver with a different port (not to connect to
the database; if ij can't, another network server cannot either).
Have you tried booting your system since?


View raw message