[OSM-dev] osm2pgsql Help ?

[Ian Dees]

...which (if the transaction was rolled back) could be vacuumed away next
time you run vacuum manually or the autovacuum process decides to run.

On Tue, Oct 12, 2010 at 10:32 AM, Ronan Crowley <ronan.crowley at gmail.com>wrote:

> I don't think so .... 116GB of tablespace files have been created on my
> /mnt/sdo drive.
>
>
>
> On 12 October 2010 16:28, Ian Dees <ian.dees at gmail.com> wrote:
>
>> On Tue, Oct 12, 2010 at 9:32 AM, Ronan Crowley <ronan.crowley at gmail.com>wrote:
>>
>>> Hi,
>>>
>>> I have just run the osm2pgsql script.
>>> It has taken over 4 days for me.
>>>
>>> However, the user I imported into had a tablespace that was on  /mnt/sdo
>>>
>>> When the program got to building the indexes, I got the following error :
>>>
>>> CREATE INDEX planet_osm_ways_nodes ON planet_osm_ways USING gin (nodes
>>> gin__int_ops) TABLESPACE pg_default;
>>>  failed: ERROR:  could not extend relation base/16391/832804: No space
>>> left on device
>>> HINT:  Check free disk space.
>>>
>>> Now looking at the code I can see that had I used a -i parameter I could
>>> have avoided this.
>>>
>>> What I want to know is, is there any way I can manually continue the
>>> outstanding steps  ?
>>> (Perhaps redoing the )
>>> OR
>>> Is there a whay to re-run the program, just to create the indexes ?
>>>
>>> Thanks,
>>> Ronan
>>>
>>>
>> I might be wrong, but I think the osm2pgsql actions occur within a
>> transaction. Since your transaction failed it most likely has been rolled
>> back, thus leaving you with an essentially empty database. People more
>> familiar with osm2pgsql should tell me if I'm wrong, though.
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/dev/attachments/20101012/fedc5684/attachment.html>