summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2010-08-12failover to reading dbs.couch directly when ets is MIAAdam Kocoloski
2010-08-12fix debugging funs and clean up API moduleAdam Kocoloski
2010-08-12mega refactoring of mem3Adam Kocoloski
2010-08-12use name_shard/1 instead of obsolete shard_name/2Adam Kocoloski
2010-08-12converter from Doc to #shard{} is more generalAdam Kocoloski
2010-08-12rewrite mem3_cache to use continuous _changes feedAdam Kocoloski
2010-08-12move states/0 to top-level moduleAdam Kocoloski
2010-08-12simpler event handler for sync serverAdam Kocoloski
2010-08-12update test_suiteAdam Kocoloski
2010-08-12better naming for event manager and sync event handlerAdam Kocoloski
2010-08-12remove unused/obsolete test codeAdam Kocoloski
2010-08-12update to new vclock module nameAdam Kocoloski
2010-08-12standardize mem3 naming. app is horribly broken for nowAdam Kocoloski
2010-08-12clean up membership application callback modAdam Kocoloski
2010-08-12add distinct membership http req handler to membership app, not showroomBrad Anderson
2010-08-12patch up bugzid 10333, a cast on nodeup now sends gossip as wellBrad Anderson
2010-08-12add membership events back into mem3, and add dbs handlersBrad Anderson
2010-08-12skip non-membership docs in dbs.couchAdam Kocoloski
2010-08-12atoms are not ejson keys, use bins for http compatibilityAdam Kocoloski
2010-08-12better handling of repeated create_db attempts. BugzID 10300Adam Kocoloski
2010-08-12mem3:nodes() does the trick hereAdam Kocoloski
2010-08-12remove hash_int(), just convert hash() to int inlineAdam Kocoloski
2010-08-12remove unused codeAdam Kocoloski
2010-08-12more detailed report for mem3:states()Adam Kocoloski
2010-08-12faster+simpler get_pingnode_stateAdam Kocoloski
2010-08-12missed a lib_misc function that vector clocks used, during dynomite -> ↵Brad Anderson
membership rename
2010-08-12dbs database creation moved to membership appBrad Anderson
2010-08-12throw an error in all_parts if DB does not existAdam Kocoloski
2010-08-12forgot to start the db childrenBrad Anderson
2010-05-28begin move of dynomite to membershipBrad Anderson
2010-05-28remove dynomite cruftBrad Anderson
2010-05-27open_doc call in fabricAdam Kocoloski
2010-05-27allow for single-node cluster 'init'Brad Anderson
2010-05-27replace #part with #shardAdam Kocoloski
2010-05-26change {N,P} fullmap over to #part{} record for future partition-splitting ↵Brad Anderson
hotness
2010-05-25all_databases now working with/without Customer param, types moved to ↵Brad Anderson
mem.hrl, view updater thwarted for updates to dbs db.
2010-05-24http admin call for join 'replace'Brad Anderson
2010-05-24commentsBrad Anderson
2010-05-24some mem3 docs/specsBrad Anderson
2010-05-20add another clause for join-replaceBrad Anderson
2010-05-19mem3 code for node replacementBrad Anderson
2010-05-19remove all vector clocks from #doc, and code to go with it. Not sure if ↵Brad Anderson
things still work, but are being rewritten anyway with mem3. Closes BugzID 9959
2010-05-14removing more of dynomite deps from mem3Brad Anderson
2010-05-12create_db quorum check done, as well as initial install_fullmap functionBrad Anderson
2010-05-11move partitions into showroom, now that there's not much left of dynomite codeBrad Anderson
2010-05-09undo last commit, let's leave cluster_ops alone, make a new showroom moduleBrad Anderson
2010-05-09change cluster_ops:all_parts to accept the list of NodePartsBrad Anderson
2010-05-09work on create_db functionality, can now generate fullmap for a db based on ↵Brad Anderson
its name, its config, and current mem3 nodes - BugzID 10007
2010-05-09disable some partition calls, and results from http admin layerBrad Anderson
2010-05-09add 0.5sec sleep to ping_all, seems to have solved 'init' bug picked up by ↵Brad Anderson
integration tester
generated by cgit v1.2.3 (git 2.25.1) at 2025-08-15 20:53:14 +0000