OK, thanks for this hint! I set it now to: log > level > debug
I run a single couchdb instance Yes, this could be a timeout error, when Wikipedia maybe stops queries from my servers ip, thinking it is too much. But this is only a guess. When I re-run the scrape, it works immediately again. The scrape with the 500 error can happen after some hours, or nearly a day. Anyway I put the timeout of [couch_httpd_auth] now to timeout = 3600 and will see, if it something changes. Thanks! frank Am 03.05.2020 um 13:04 schrieb Ronny Berndt: > You should increase your log level to debug to get more infos. > > Do you run a cluster or a single couchdb instance? > You said that you are getting an error „after a while“. Is this time > in the same time frame - e.g. a timeout error? > > > > >> Am 03.05.2020 um 12:53 schrieb Frank Röhm <[email protected]>: >> >> Just this: >> >> couchdb.http.ServerError: (500, (u'unknown_error', u'undefined')) >> >> >>> Am 03.05.2020 um 12:44 schrieb Ronny Berndt <[email protected]>: >>> >>> Are there any error details in the CouchDB log file? >>> >>> Ronny >>> >>>> Am 03.05.2020 um 11:42 schrieb Frank Röhm <[email protected]>: >>>> >>>> Hallo >>>> >>>> is it possible to downgrade on a Ubuntu 18.04 server to the old CouchDB >>>> 1.6? >>>> >>>> I have since CouchDB 2.x problems while filling a db with Wikipedia >>>> entries. I use therefore mwscrape from: >>>> https://github.com/itkach/mwscrape >>>> and after a while I always get a 500 error from CouchDB and my scrape >>>> stops. This never happened with CouchDB 1.6 under Ubuntu 14.04. >>>> >>>> So could it be possible to install CouchDB 1.x maybe by self-compilation >>>> somehow or are there old and now unsatisfiable dependencies? Someone knows? >>>> >>>> Thanks, frank >>>> >> >
