Age | Commit message (Collapse) | Author |
|
Our headers start with a <<1>> and then four bytes indicating the length
of the header and its checksum. When the header is larger than 4090
bytes it will be split across multiple blocks in the file and will need
to be reassembled on read. The reassembly consists of stripping out
<<0>> from the beginning of each subsequent block in the
remove_block_prefixes/2 function. The bug here is that we tell
remove_block_prefixes that we're starting 1 byte into the current block
instead of 5, so it ends up removing one good byte from the header and
injecting one or more random <<0>>s.
Headers larger than 4k are very rare and generally require a view group
with a huge number of indexes or indexes with fairly large reductions,
which explains why this bug has gone undetected until now.
Closes COUCHDB-1319.
|
|
|
|
BugzID:12741
|
|
The race condition in couch_server's ets table usage rears its ugly head
by leaving an entry in couch_lru. This patch just addresses the issue by
allowing the client pid to use the db and ignores the fact that for the
duration its over the max_dbs_open setting.
|
|
|
|
|
|
|
|
I was slaving away on this merge for two days.
|
|
|
|
Conflicts:
apps/couch/include/couch_db.hrl
apps/couch/src/couch_db.erl
apps/couch/src/couch_os_process.erl
apps/couch/src/couch_query_servers.erl
apps/couch/src/couch_rep.erl
apps/couch/src/couch_replication_manager.erl
apps/couch/src/couch_view_compactor.erl
apps/couch/src/couch_view_group.erl
apps/couch/src/couch_view_updater.erl
configure.ac
couchjs/c_src/http.c
couchjs/c_src/main.c
couchjs/c_src/utf8.c
etc/windows/couchdb.iss.tpl
src/couchdb/priv/Makefile.am
src/couchdb/priv/couch_js/main.c
test/etap/160-vhosts.t
test/etap/200-view-group-no-db-leaks.t
test/etap/Makefile.am
BugzID: 12645
|
|
The `created` event is emitted on apache couchdb when a database is
created. This patch re-add it to bigcouch.
|
|
|
|
|
|
1) couch_file:sync could leave open fd's if close failed. Now we'll get a trace.
2) couch_file:append_term failing would be bad, so let's test that too.
backported from trunk r1150915
git-svn-id: https://svn.apache.org/repos/asf/couchdb/branches/1.1.x@1150918 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
git-svn-id: https://svn.apache.org/repos/asf/couchdb/branches/1.1.x@1142410 13f79535-47bb-0310-9956-ffa450edef68
|
|
On server startup, restart replications in error
If we setup a continuous replication which goes into
an error state and restart Couch just before the
replication is retried (before it transitions to the
triggered state), the user has to manually restart
the replication (recreating the document or deleting
its _replication_state field).
git-svn-id: https://svn.apache.org/repos/asf/couchdb/branches/1.1.x@1142263 13f79535-47bb-0310-9956-ffa450edef68
|
|
Replication manager, ignore db monitor messages
Ignore local database monitor messages. These messages
are received in some circunstances because replication
processes spawn local database monitors but never
do the corresponding erlang:demonitor/1,2 calls.
git-svn-id: https://svn.apache.org/repos/asf/couchdb/branches/1.1.x@1142260 13f79535-47bb-0310-9956-ffa450edef68
|
|
If a replication transitions to the "error" state, attempt
to restart it up to "max_replication_retry_count" times
(.ini configuration parameter). This number of retry attempts
can now be set to "infinity" as well.
This was already current behaviour in trunk (upcoming 1.2).
Closes COUCHDB-1194.
git-svn-id: https://svn.apache.org/repos/asf/couchdb/branches/1.1.x@1142258 13f79535-47bb-0310-9956-ffa450edef68
Conflicts:
apps/couch/src/couch_rep.erl
|
|
Improved error message in the replicator
git-svn-id: https://svn.apache.org/repos/asf/couchdb/branches/1.1.x@1140887 13f79535-47bb-0310-9956-ffa450edef68
|
|
Add infinity timeout to couch_ref_counter calls
After compacting a very large database, the updater calls the couch_db gen_server with
a db record that contains a new ref counter. The couch_db gen_server calls drop on the
old ref counter and calls add on the new ref counter. However since the system is busy
deleting the old db file or garbage collecting, one of the ref counter calls times out,
causing couch_db's terminate to invoked and terminate calls shutdown on the updater.
However the updater is waiting for the call it made to couch_db to complete, which can't
complete since it's waiting for the updater.
git-svn-id: https://svn.apache.org/repos/asf/couchdb/branches/1.1.x@1138799 13f79535-47bb-0310-9956-ffa450edef68
|
|
BugzID: 12365
|
|
Conflicts:
acinclude.m4.in
apps/couch/src/couch_app.erl
apps/couch/src/couch_doc.erl
apps/couch/src/couch_view.erl
configure.ac
share/www/script/test/replicator_db.js
src/erlang-oauth/Makefile.am
test/etap/Makefile.am
|
|
|
|
|
|
|
|
Thanks Hardtke for ruining my afternoon^H^H^H^H demonstrating the bug.
BugzID: 12112
|
|
|
|
The #full_doc_info record is extended to include the summed size of
leaf revision document bodies and their attachments. Document sizes
are computed on update; accurate sizes of existing databases and
view groups are only available after compaction.
The document size is defined to be the size of the binary
representation of #doc.body. The att_len field is used for
attachments; attachments that are shared by multiple revisions of a
document are only counted once. The size of a view index is defined as
the size of all keys, values, and reductions accessible from the current
root of the tree.
BugzID: 9995
|
|
This function can be used by many of the existing functions
in couch_key_tree that need to walk the revision tree, apply
a function at each branch or leaf, and then accumulate results.
|
|
This patch has the side effect of fixing a crash of the _users DB after
the first minute of running. The crash occurs because BigCouch presumes
that couch_stats_collector is monitoring the file descriptor, but
nothing prevents the supervisor from starting couch_auth_cache (and thus
opening the _users DB) before couch_stats_collector is started. The
start order is determined by the ordering of the children from the
config file in couch_config's ets set.
Closes #39
BugzID: 11854
|
|
BugzID: 12043
|
|
|
|
Thanks henrik14091, mc_ for the report
|
|
|
|
|
|
BugzID: 11498
|
|
This patch causes the view_updater to wait until either a minimum
number of items or a maximum process memory size has been reached
before committing the changes to disk.
See COUCHDB-700 for discussion
BugzID: 11885
|
|
|
|
|
|
BugzID: 11797
|
|
|
|
This reverts commit d2a9229c4f0a750d8e4191f60227572117a53ab7.
|
|
|
|
git-svn-id: https://svn.apache.org/repos/asf/couchdb/branches/1.0.x@1064800 13f79535-47bb-0310-9956-ffa450edef68
|
|
This reverts commit 5c3161d57b1d2d7953497c24be09112b6bcf71b6.
|
|
|
|
|