You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@couchdb.apache.org by Sinan Gabel <si...@gmail.com> on 2017/01/05 14:29:54 UTC
Re: Replication breaks to target on localhost single-node 2.0 install
Hi!
The problem I experienced with replication break, using CouchDB 2.0
single-node on Desktop against CouchDB 1.x.x. through nginx on cloud, is
solved when upgrading to CouchDB 2.0 cluster through nginx on cloud.
That is, the problem has been solved in CouchDB 2.0, although I do not know
the cause and resolve. Anyway, case closed!
Thanks!
PS I have made a full description of the CouchDB 2.0 Cluster setup I am
using on Ubuntu 16.04 servers:
https://gist.github.com/SinanGabel/eac83a2f9d0ac64e2c9d4bd936be9313
On 22 December 2016 at 12:50, Sinan Gabel <si...@gmail.com> wrote:
> Hi!
>
> I am testing couchdb 2.x (version commit 20 Dec. 2016), and have made a
> single node install on Ubuntu Desktop 16.04.
>
> I also have a 1.x.x (later than 1.6.1 but not the big couchdb version)
> cloud-based pair of couchdb's with nginx ssl load balancer in front of the
> two cloud-based couchdb's. There is a domain name "riskbutler.net"
> attached to the nginx load balancer. These two cloud-based couchdb's work
> without any problems.
>
> (a) Replicating a 85 MB database with 29.000 documents works fine from
> *source*: https://user:password@*riskbutler.net <http://riskbutler.net>*/database
> to *target:* "database" on Mac desktop with couchdb *1.6.1* on localhost.
> Thus OK here!
>
> (b) Same database replication also works fine if specifying *source: *
> https://user:password@*123.456.789.123*/database i.e. using the exact IP
> address of *one* of the two cloud-based couchdb's i.e. not going through
> the load balancer to *target:* "database" on Ubuntu Desktop 16.04 couchdb
> *2.x*. Thus OK here!
>
> (c) PROBLEM HERE: However, doing *source*: https://user:password@*riskbutler.net
> <http://riskbutler.net>*/database to *target:* "database" or *target:* "
> http://user:password@localhost:5984/database" on Ubuntu Desktop 16.04
> couchdb *2.x*. the replication breaks with message:
>
> --- extract ---
> ...
> [notice] 2016-12-22T10:25:04.195883Z couchdb@localhost <0.4261.3>
> 2b8b893bbd localhost:5984 127.0.0.1 user POST /test5/_bulk_docs 201 ok 188
> [notice] 2016-12-22T10:25:06.251197Z couchdb@localhost <0.4261.3>
> 6093a132cf localhost:5984 127.0.0.1 user POST /test5/_bulk_docs 201 ok 176
> [error] 2016-12-22T10:25:06.273789Z couchdb@localhost <0.16438.5>
> -------- Replicator, request GET to "https://user:*****@riskbutler
> .net/database/databaseFXhour1469755505?revs=true&open_revs=%5B%221-
> 51d946f83ab8d372e3a24bf8f0667abb%22%5D&latest=true" failed due to error
> {error,{conn_failed,{error,nxdomain}}}
> [notice] 2016-12-22T10:25:06.273993Z couchdb@localhost <0.16436.5>
> -------- Retrying GET to https://user:*****@riskbutler.net/database/
> databaseFXhour1469755505?revs=true&open_revs=%5B%221-
> 51d946f83ab8d372e3a24bf8f0667abb%22%5D&latest=true in 1.0 seconds due to
> error {error,{error,{conn_failed,{error,nxdomain}}}}
> [error] 2016-12-22T10:25:06.274578Z couchdb@localhost <0.16445.5>
> -------- Replicator, request GET to "https://user:*****@riskbutler
> .net/database/databaseFXhour1471680058?revs=true&open_revs=%5B%221-
> 63cb249a19f0a78f7ac7a5bc02c3d268%22%5D&latest=true" failed due to error
> {error,{conn_failed,{error,nxdomain}}}
> [notice] 2016-12-22T10:25:06.274695Z couchdb@localhost <0.16443.5>
> -------- Retrying GET to https://user:*****@riskbutler.net/database/
> databaseFXhour1471680058?revs=true&open_revs=%5B%221-
> 63cb249a19f0a78f7ac7a5bc02c3d268%22%5D&latest=true in 1.0 seconds due to
> error {error,{error,{conn_failed,{error,nxdomain}}}}
> [notice] 2016-12-22T10:25:07.689150Z couchdb@localhost <0.4261.3>
> 6a9349d77c localhost:5984 127.0.0.1 user POST /test5/_bulk_docs 201 ok 107
> [notice] 2016-12-22T10:25:07.715061Z couchdb@localhost <0.4261.3>
> 31a5201569 localhost:5984 127.0.0.1 user POST /test5/_revs_diff 200 ok 19
> ...
> --- end extract ---
>
> I have tried with and without use_checkpoints in the configuration, no
> difference in the result.
>
> Any ideas for testing/solving this are very welcome!
>
> Thanks,
>
> Br,
> Sinan
>
Re: Replication breaks to target on localhost single-node 2.0 install
Posted by Sinan Gabel <si...@gmail.com>.
Thanks, Adam!
On 5 January 2017 at 15:40, Adam Kocoloski <ko...@apache.org> wrote:
> Hi Sinan, glad you got it sorted out. For reference, the {error, nxdomain}
> in the logs means CouchDB couldn’t resolve the DNS name in order to submit
> the HTTP GET request. Cheers,
>
> Adam
>
> > On Jan 5, 2017, at 9:29 AM, Sinan Gabel <si...@gmail.com> wrote:
> >
> > Hi!
> >
> > The problem I experienced with replication break, using CouchDB 2.0
> > single-node on Desktop against CouchDB 1.x.x. through nginx on cloud, is
> > solved when upgrading to CouchDB 2.0 cluster through nginx on cloud.
> >
> > That is, the problem has been solved in CouchDB 2.0, although I do not
> know
> > the cause and resolve. Anyway, case closed!
> >
> > Thanks!
> >
> > PS I have made a full description of the CouchDB 2.0 Cluster setup I am
> > using on Ubuntu 16.04 servers:
> > https://gist.github.com/SinanGabel/eac83a2f9d0ac64e2c9d4bd936be9313
> >
> >
> > On 22 December 2016 at 12:50, Sinan Gabel <si...@gmail.com> wrote:
> >
> >> Hi!
> >>
> >> I am testing couchdb 2.x (version commit 20 Dec. 2016), and have made a
> >> single node install on Ubuntu Desktop 16.04.
> >>
> >> I also have a 1.x.x (later than 1.6.1 but not the big couchdb version)
> >> cloud-based pair of couchdb's with nginx ssl load balancer in front of
> the
> >> two cloud-based couchdb's. There is a domain name "riskbutler.net"
> >> attached to the nginx load balancer. These two cloud-based couchdb's
> work
> >> without any problems.
> >>
> >> (a) Replicating a 85 MB database with 29.000 documents works fine from
> >> *source*: https://user:password@*riskbutler.net <http://riskbutler.net
> >*/database
> >> to *target:* "database" on Mac desktop with couchdb *1.6.1* on
> localhost.
> >> Thus OK here!
> >>
> >> (b) Same database replication also works fine if specifying *source: *
> >> https://user:password@*123.456.789.123*/database i.e. using the exact
> IP
> >> address of *one* of the two cloud-based couchdb's i.e. not going through
> >> the load balancer to *target:* "database" on Ubuntu Desktop 16.04
> couchdb
> >> *2.x*. Thus OK here!
> >>
> >> (c) PROBLEM HERE: However, doing *source*: https://user:password@*
> riskbutler.net
> >> <http://riskbutler.net>*/database to *target:* "database" or *target:*
> "
> >> http://user:password@localhost:5984/database" on Ubuntu Desktop 16.04
> >> couchdb *2.x*. the replication breaks with message:
> >>
> >> --- extract ---
> >> ...
> >> [notice] 2016-12-22T10:25:04.195883Z couchdb@localhost <0.4261.3>
> >> 2b8b893bbd localhost:5984 127.0.0.1 user POST /test5/_bulk_docs 201 ok
> 188
> >> [notice] 2016-12-22T10:25:06.251197Z couchdb@localhost <0.4261.3>
> >> 6093a132cf localhost:5984 127.0.0.1 user POST /test5/_bulk_docs 201 ok
> 176
> >> [error] 2016-12-22T10:25:06.273789Z couchdb@localhost <0.16438.5>
> >> -------- Replicator, request GET to "https://user:*****@riskbutler
> >> .net/database/databaseFXhour1469755505?revs=true&open_revs=%5B%221-
> >> 51d946f83ab8d372e3a24bf8f0667abb%22%5D&latest=true" failed due to error
> >> {error,{conn_failed,{error,nxdomain}}}
> >> [notice] 2016-12-22T10:25:06.273993Z couchdb@localhost <0.16436.5>
> >> -------- Retrying GET to https://user:*****@riskbutler.net/database/
> >> databaseFXhour1469755505?revs=true&open_revs=%5B%221-
> >> 51d946f83ab8d372e3a24bf8f0667abb%22%5D&latest=true in 1.0 seconds due
> to
> >> error {error,{error,{conn_failed,{error,nxdomain}}}}
> >> [error] 2016-12-22T10:25:06.274578Z couchdb@localhost <0.16445.5>
> >> -------- Replicator, request GET to "https://user:*****@riskbutler
> >> .net/database/databaseFXhour1471680058?revs=true&open_revs=%5B%221-
> >> 63cb249a19f0a78f7ac7a5bc02c3d268%22%5D&latest=true" failed due to error
> >> {error,{conn_failed,{error,nxdomain}}}
> >> [notice] 2016-12-22T10:25:06.274695Z couchdb@localhost <0.16443.5>
> >> -------- Retrying GET to https://user:*****@riskbutler.net/database/
> >> databaseFXhour1471680058?revs=true&open_revs=%5B%221-
> >> 63cb249a19f0a78f7ac7a5bc02c3d268%22%5D&latest=true in 1.0 seconds due
> to
> >> error {error,{error,{conn_failed,{error,nxdomain}}}}
> >> [notice] 2016-12-22T10:25:07.689150Z couchdb@localhost <0.4261.3>
> >> 6a9349d77c localhost:5984 127.0.0.1 user POST /test5/_bulk_docs 201 ok
> 107
> >> [notice] 2016-12-22T10:25:07.715061Z couchdb@localhost <0.4261.3>
> >> 31a5201569 localhost:5984 127.0.0.1 user POST /test5/_revs_diff 200 ok
> 19
> >> ...
> >> --- end extract ---
> >>
> >> I have tried with and without use_checkpoints in the configuration, no
> >> difference in the result.
> >>
> >> Any ideas for testing/solving this are very welcome!
> >>
> >> Thanks,
> >>
> >> Br,
> >> Sinan
> >>
>
>
Re: Replication breaks to target on localhost single-node 2.0 install
Posted by Adam Kocoloski <ko...@apache.org>.
Hi Sinan, glad you got it sorted out. For reference, the {error, nxdomain} in the logs means CouchDB couldn’t resolve the DNS name in order to submit the HTTP GET request. Cheers,
Adam
> On Jan 5, 2017, at 9:29 AM, Sinan Gabel <si...@gmail.com> wrote:
>
> Hi!
>
> The problem I experienced with replication break, using CouchDB 2.0
> single-node on Desktop against CouchDB 1.x.x. through nginx on cloud, is
> solved when upgrading to CouchDB 2.0 cluster through nginx on cloud.
>
> That is, the problem has been solved in CouchDB 2.0, although I do not know
> the cause and resolve. Anyway, case closed!
>
> Thanks!
>
> PS I have made a full description of the CouchDB 2.0 Cluster setup I am
> using on Ubuntu 16.04 servers:
> https://gist.github.com/SinanGabel/eac83a2f9d0ac64e2c9d4bd936be9313
>
>
> On 22 December 2016 at 12:50, Sinan Gabel <si...@gmail.com> wrote:
>
>> Hi!
>>
>> I am testing couchdb 2.x (version commit 20 Dec. 2016), and have made a
>> single node install on Ubuntu Desktop 16.04.
>>
>> I also have a 1.x.x (later than 1.6.1 but not the big couchdb version)
>> cloud-based pair of couchdb's with nginx ssl load balancer in front of the
>> two cloud-based couchdb's. There is a domain name "riskbutler.net"
>> attached to the nginx load balancer. These two cloud-based couchdb's work
>> without any problems.
>>
>> (a) Replicating a 85 MB database with 29.000 documents works fine from
>> *source*: https://user:password@*riskbutler.net <http://riskbutler.net>*/database
>> to *target:* "database" on Mac desktop with couchdb *1.6.1* on localhost.
>> Thus OK here!
>>
>> (b) Same database replication also works fine if specifying *source: *
>> https://user:password@*123.456.789.123*/database i.e. using the exact IP
>> address of *one* of the two cloud-based couchdb's i.e. not going through
>> the load balancer to *target:* "database" on Ubuntu Desktop 16.04 couchdb
>> *2.x*. Thus OK here!
>>
>> (c) PROBLEM HERE: However, doing *source*: https://user:password@*riskbutler.net
>> <http://riskbutler.net>*/database to *target:* "database" or *target:* "
>> http://user:password@localhost:5984/database" on Ubuntu Desktop 16.04
>> couchdb *2.x*. the replication breaks with message:
>>
>> --- extract ---
>> ...
>> [notice] 2016-12-22T10:25:04.195883Z couchdb@localhost <0.4261.3>
>> 2b8b893bbd localhost:5984 127.0.0.1 user POST /test5/_bulk_docs 201 ok 188
>> [notice] 2016-12-22T10:25:06.251197Z couchdb@localhost <0.4261.3>
>> 6093a132cf localhost:5984 127.0.0.1 user POST /test5/_bulk_docs 201 ok 176
>> [error] 2016-12-22T10:25:06.273789Z couchdb@localhost <0.16438.5>
>> -------- Replicator, request GET to "https://user:*****@riskbutler
>> .net/database/databaseFXhour1469755505?revs=true&open_revs=%5B%221-
>> 51d946f83ab8d372e3a24bf8f0667abb%22%5D&latest=true" failed due to error
>> {error,{conn_failed,{error,nxdomain}}}
>> [notice] 2016-12-22T10:25:06.273993Z couchdb@localhost <0.16436.5>
>> -------- Retrying GET to https://user:*****@riskbutler.net/database/
>> databaseFXhour1469755505?revs=true&open_revs=%5B%221-
>> 51d946f83ab8d372e3a24bf8f0667abb%22%5D&latest=true in 1.0 seconds due to
>> error {error,{error,{conn_failed,{error,nxdomain}}}}
>> [error] 2016-12-22T10:25:06.274578Z couchdb@localhost <0.16445.5>
>> -------- Replicator, request GET to "https://user:*****@riskbutler
>> .net/database/databaseFXhour1471680058?revs=true&open_revs=%5B%221-
>> 63cb249a19f0a78f7ac7a5bc02c3d268%22%5D&latest=true" failed due to error
>> {error,{conn_failed,{error,nxdomain}}}
>> [notice] 2016-12-22T10:25:06.274695Z couchdb@localhost <0.16443.5>
>> -------- Retrying GET to https://user:*****@riskbutler.net/database/
>> databaseFXhour1471680058?revs=true&open_revs=%5B%221-
>> 63cb249a19f0a78f7ac7a5bc02c3d268%22%5D&latest=true in 1.0 seconds due to
>> error {error,{error,{conn_failed,{error,nxdomain}}}}
>> [notice] 2016-12-22T10:25:07.689150Z couchdb@localhost <0.4261.3>
>> 6a9349d77c localhost:5984 127.0.0.1 user POST /test5/_bulk_docs 201 ok 107
>> [notice] 2016-12-22T10:25:07.715061Z couchdb@localhost <0.4261.3>
>> 31a5201569 localhost:5984 127.0.0.1 user POST /test5/_revs_diff 200 ok 19
>> ...
>> --- end extract ---
>>
>> I have tried with and without use_checkpoints in the configuration, no
>> difference in the result.
>>
>> Any ideas for testing/solving this are very welcome!
>>
>> Thanks,
>>
>> Br,
>> Sinan
>>