Commit graph

246 commits

Author SHA1 Message Date
Travis
345d076fdf
introduce "fields" directory between index and field 2021-03-05 16:56:44 -06:00
Travis
4e7ec34c6d
change directory ".disco" to "disco" 2021-03-05 16:56:44 -06:00
Travis
7789e24965
introduce "indexes" directory between datadir and index 2021-03-05 16:56:43 -06:00
Nia
d205ea5441
Merge pull request #1442 from niaow/fix-foreignindex-race
Fix a race condition when deferring foreign-index initialization
2021-02-26 14:30:18 -05:00
Nia Weiss
3fbf6993d0
defer cluster messages until startup 2021-02-25 16:01:49 -05:00
Nia Weiss
ca918c187d
fix a race condition when deferring foreign-index initialization 2021-02-25 16:00:56 -05:00
Travis
2bbe1fdde0
remove remaining references to "coordinator" 2021-02-23 17:23:09 -06:00
Travis
8b0f18721e
remove Field.loadMeta() 2021-02-23 10:09:55 -06:00
Travis
d639e228ae
remove Index.saveMeta(). remove support for deleing existence field. 2021-02-23 10:09:55 -06:00
Travis
38d5459e25
convert holder decode* methods to functions 2021-02-23 10:09:55 -06:00
Travis
24c5b654c6
change IndexOptions to reference by value 2021-02-23 10:09:54 -06:00
Travis
e54f7d9a0b
ensure field.CreatedAt is set on loadField 2021-02-15 21:28:20 -06:00
Travis
b5d6c632bf
stop storing a value for views in etcd 2021-02-15 14:39:52 -06:00
Seebs
e2b6912d1c uninvert test for coordinator node / primary field translation node
If we're the primary field translation node, we don't need to set up
translation replication; we only need that if we're *not*. So it
makes sense to test if !IsPrimaryFieldTranslationNode... except that
the test is to determine whether to return early. So it should not
be inverted.
2021-02-15 10:19:03 -06:00
Travis
8f0270acda
adjust openExistenceField() to check on disk first 2021-02-12 20:35:36 -06:00
Kuba Podgórski
2f35b51db8
Fix endpoint tests + change BitDepth type to uint64 2021-02-12 20:35:36 -06:00
Travis
a6297dc48e
WIP: load schema from etcd on holder open; validate indexes, fields, views 2021-02-12 20:32:31 -06:00
Travis
c37424f24f
remove commented code 2021-02-10 11:57:23 -06:00
Travis
1c3b0f364d
implement ApplySchema, LoadSchema, and LoadSchemaMessage 2021-02-10 11:51:52 -06:00
Travis
bcb71b023a
fix misplaced _exists check 2021-02-10 11:47:26 -06:00
Travis
8fe1b9a37c
store views in etcd via Schemator 2021-02-10 11:47:26 -06:00
Kuba Podgórski
9faabd8535 Fix Test_TxFactory_UpdateBlueFromGreen_OnStartup 2021-02-10 14:26:52 +01:00
Kuba Podgórski
40063a6aa9 Fix TestHandler_PostSchemaCluster 2021-02-09 20:13:38 +01:00
Kuba Podgórski
0f97066914 Fix TestAPI_Import. Delete views when deleting a field. 2021-02-09 13:36:53 +01:00
Kuba Podgórski
37b309be16 Move schemator from API to index. Fix TestExecutor_Execute_SetRow 2021-02-08 21:54:39 +01:00
Antonio Navarro Perez
17a5bc51ca Add NopSchemator and NopSerializer
Signed-off-by: Antonio Navarro Perez <antnavper@gmail.com>
2021-02-08 18:36:53 +01:00
Travis
3542134100
WIP: implement Schemator 2021-02-08 10:42:55 -06:00
Travis
114f6a8751
add withViews argument to api.Schema() method 2021-02-08 10:42:55 -06:00
Travis
d192c1f24f
Merge branch 'master' into disco 2021-02-05 15:58:36 -06:00
Travis
ec91dad198
remove dead code 2021-02-05 11:31:33 -06:00
Travis
da804ee6d5
linter fixes 2021-02-03 15:11:24 -06:00
Antonio Navarro Perez
c45e21640c
Change coordinator to primary
Signed-off-by: Antonio Navarro Perez <antnavper@gmail.com>
2021-02-02 15:56:12 -06:00
Travis
855e1b35f5
more use of noder; remove c.nodes
disable some of the gossip logic

implement some of the stator logic
2021-01-31 23:42:49 -06:00
Alan Bernstein
15443b371c Force consistent timestamp width in startup log 2021-01-29 17:48:24 -06:00
Travis
b80f5099b2
more coordinator/primary cleanup 2021-01-25 23:20:08 -06:00
Travis
ace4dea46f
address some test failures due to random ordered etcd ID 2021-01-25 00:52:49 -06:00
Travis
2f66501160
finish implementing snap := ClusterSnapshot() 2021-01-24 23:22:32 -06:00
Travis
19f91782e7
remove all instances of txsrc 2021-01-21 21:59:51 -06:00
Travis
13984353e4
remove instances of os.Getenv("PILOSA_TXSRC") 2021-01-21 21:18:46 -06:00
Travis
08fae2be4c
introduce storage.Config 2021-01-20 22:05:38 -06:00
Travis
bc13834343
disco/etcd work: fix lots of races, start all cluster nodes at once.
port mapper gives out ports from 63000-65000 for the tests

fix another race

http test uses port.MustGetPort

rbf: remove :0 port request

ocd happy

test fix for grpc listener address already in use

test/disco allocates BindGRPC port from the port mapper

dump stack on each GetPort

verify each port is usable right away

server/config.go has Config.Validate() now

panic if gossip port is 0. validate server.Config

fix another gossip port 0

builds

quiet, don't dump stack on each port alloc

builds

happy linter

even gossip fallback should not be zero but rather use the port mapper
2021-01-12 21:06:12 -06:00
Travis
3f26d667b4
remove pilosa.DefaultPartitionN 2021-01-06 22:45:44 -06:00
Travis
4515a24e48
change all references to use subpackages: topology, net 2021-01-06 16:09:24 -06:00
Jason E. Aten
035073555a pilosa: only open views with data
- Previously, on timequantum schemas, we would
create and open a view for the cartesian
product of every possible view and shard.

- This caused us to be very slow on re-open,
and to use lots of memory for views that
held nothing.

- This change makes startup faster, memory
use much lower, and should speed migration.
2020-12-19 00:03:31 +00:00
Nia Weiss
eab70e2314
use the txf to determine if snapshots are needed 2020-12-16 11:00:27 -05:00
Nia Weiss
aaf94c6aeb
fix the automatic snapshot queue enable check
The "needs snapshot queue" check was broken, as it only checked inside a loop over indices.
If there are no indexes yet (or more likely if the indices have not yet been loaded off of disk), then this would never use the snapshot queue on roaring.
2020-12-16 08:15:17 -05:00
Jason E. Aten
b254c6776b FragSpec as struct, fragment fields to avoid dynamic dispatch
- allow inlining of getters index(), view().
 - GOMAXPROCS set to 128
2020-12-12 16:39:24 +00:00
Jason E. Aten
03a54c6d5f rowcache off by default. pilosa server --rowcache-on turns it back on. 2020-12-08 23:18:45 +00:00
Nia Weiss
7ea2e4b7e3
add ID auto-generation on the coordinator 2020-12-03 13:04:14 -05:00
Jason E. Aten
7e22994a6d run migration in parallel
- migration can be slow. parallelize it.
2020-11-05 18:31:27 +00:00