Commit Graph

750 Commits

Author SHA1 Message Date
Łukasz Magiera
94009f247f
Merge pull request from filecoin-project/feat/sealing-handle-batch-publish
Handle batch publish storage deals message in sealing recovery
2021-01-28 19:33:49 +01:00
Dirk McCormick
1b494acc9e feat: handle batch publish storage deals message in sealing recovery 2021-01-25 14:10:05 +01:00
Steven Allen
8de4e5e005 retry proofs if we generate an incorrect one 2021-01-22 13:31:47 -08:00
Aayush Rajasekaran
ab90a3b2bc Actors update: MinerInfo.SealProofType has been removed 2021-01-21 15:21:19 -08:00
Jack Yao
41e9345da0
Update storage/addresses.go
Co-authored-by: Jakub Sztandera <kubuxu@protonmail.ch>
2021-01-19 09:36:35 +08:00
yaohcn
10b2ec55b5 correct log in maybeUseAddress 2021-01-15 12:31:47 +08:00
Łukasz Magiera
32885e1129 termination batcher: Notify based on what was sent 2021-01-14 17:14:26 +01:00
Łukasz Magiera
144b5a1350 perning termination API 2021-01-14 12:37:23 +01:00
Łukasz Magiera
3522c8d45a SectorTerminateFlush API 2021-01-13 23:32:04 +01:00
Łukasz Magiera
174c595acc Don't declare in proving window 2021-01-13 22:19:10 +01:00
Łukasz Magiera
52cc2cd3eb Initial sector termination support 2021-01-13 00:42:01 +01:00
Jakub Sztandera
6ca5caef31
Refactor DefaultMessageSendSpec
Signed-off-by: Jakub Sztandera <kubuxu@protocol.ai>
2020-12-08 20:51:27 +01:00
Łukasz Magiera
a5a72a41eb miner: Fix control address lookup 2020-12-04 17:02:23 +01:00
Łukasz Magiera
58695dc1bc miner: fix tests 2020-12-03 12:37:11 +01:00
Łukasz Magiera
2fd93a55ac plumb AddressSelector to stoage fsm 2020-12-02 21:47:54 +01:00
Łukasz Magiera
f74a1b70ab miner: Fix actor control list cli 2020-12-02 20:46:07 +01:00
Łukasz Magiera
7246ef273f miner: Control address config for (pre)commits 2020-12-02 19:58:00 +01:00
Łukasz Magiera
95eaf13b5a sectorstorage: Fix tests 2020-12-02 00:36:32 +01:00
Łukasz Magiera
c6f85886bf miner: Add slow mode to proving check 2020-12-02 00:32:14 +01:00
Łukasz Magiera
054327b72a
wdpost: Update comment
Co-authored-by: dirkmc <dirkmdev@gmail.com>
2020-12-01 14:50:51 +01:00
Łukasz Magiera
e8d2560581 Fix tests 2020-11-27 16:47:48 +01:00
s1m0n21
663b3f4f21 add a sector check command 2020-11-26 15:02:43 +08:00
hannahhoward
218f7c2c5d feat(markets): update markets, extract commp 2020-11-20 13:29:14 -08:00
Łukasz Magiera
1999156d06 wdpost: always pick least-bad address 2020-11-19 20:59:54 +01:00
Łukasz Magiera
b4fa099257 wdpost: fix TestWDPostDoPost 2020-11-19 18:57:43 +01:00
Łukasz Magiera
b7a282decc wdpost: fix build 2020-11-19 18:30:53 +01:00
Łukasz Magiera
9f807f4aeb storageminer: Simplify AddressFor 2020-11-19 18:02:20 +01:00
Łukasz Magiera
16dbbadddf Set lower feecap on PoSt messages with low balance 2020-11-19 18:02:20 +01:00
15ho
ba56d02a1c fix: WindowPoStScheduler.checkSectors() delete useless judgment 2020-11-19 19:48:32 +08:00
Łukasz Magiera
915c5e0fdd Get most tests to pass 2020-11-16 19:03:30 +01:00
Steven Allen
05db5ce426 load proof types in window post logic
This is going to be inefficient, but it should be fine in practice.
2020-11-16 19:03:30 +01:00
Steven Allen
219c695967 fix storage adapter miner 2020-11-16 19:03:30 +01:00
Łukasz Magiera
6bea9dd178 Making sealing logic work with multiple seal proof types 2020-11-16 19:03:30 +01:00
Steven Allen
077bc83f7f explicitly abort PoSt on context cancellation 2020-10-30 14:00:54 -07:00
Steven Allen
6985af5e88 keep retrying the proof until we run out of sectors to skip
If we have a bunch of corrupted but not missing sectors on disk, we may need to
retry many times before we get a proof to pass. Simply giving up doesn't help anyone.
2020-10-30 13:21:34 -07:00
Łukasz Magiera
4100f6eead fix TestWDPostDoPost 2020-10-28 15:23:21 +01:00
Łukasz Magiera
e1da874258 Merge remote-tracking branch 'origin/master' into feat/async-restartable-workers 2020-10-27 03:31:07 +01:00
Steven Allen
0d06f8fa2b Add commands to change the worker key 2020-10-26 18:16:24 -07:00
wangchao
e65aa4cf17 worker key should change when set sender found key not equal with the value on chain 2020-10-26 15:20:09 +08:00
Łukasz Magiera
4d87473053 Fix lint 2020-10-23 23:31:18 +02:00
Łukasz Magiera
660236b224 Merge remote-tracking branch 'origin/master' into feat/async-restartable-workers 2020-10-23 23:25:35 +02:00
Steven Allen
109b326ae2 Use older randomness for the PoSt commit on specs-actors version 2
After the upgrade, we're allowed to use randomness up to the challenge epoch.
This safe-guards us against very large reorgs where a miner computes a PoSt very
quickly.
2020-10-23 12:45:11 -07:00
Łukasz Magiera
29e334de54
Merge pull request from filecoin-project/steb/generalize-window-post
Manage sectors by size instead of proof type.
2020-10-22 21:27:48 +02:00
Steven Allen
bcabe7b3b5 migrate methods to abstracted methods
Method numbers never change anyways. At worst, we'll deprecate old methods and
have to explicitly import them from the correct actors version to use them.
2020-10-21 12:18:37 -07:00
Steven Allen
4e730b5ec8 port to v2 imports 2020-10-21 12:16:23 -07:00
Steven Allen
00dcb1bce9 Manage sectors by size instead of proof type.
* We may have multiple sectors with the same size and different proof types, but all these management functions stay the same.
* This simplifies PoSt logic.
2020-10-20 18:30:56 -07:00
Łukasz Magiera
8d06cca073 sched: Handle workers using sessions instead of connections 2020-10-18 12:36:06 +02:00
Łukasz Magiera
b74a3229f8 fsm: process expired-ticket sectors 2020-10-13 21:35:41 +02:00
Jakub Sztandera
2b21fdef33
Refactor from crypto.SigType to types.KeyType
Signed-off-by: Jakub Sztandera <kubuxu@protocol.ai>
2020-10-12 00:07:53 +02:00
Steven Allen
748d2e82a7 unshare the journal
Motivation:

* Run lotus with the race detector enabled (primary motivation).
* Allow multiple lotus nodes in a process (not a high priority).

Previously, the journal was shared between all lotus instances, but it was
initialized for every new node. This caused safety problems in tests (at a
minimum).

This patch explicitly passes the journal to all services that need it.
2020-10-09 13:23:07 -07:00
Steven Allen
196bf39574 use non-faulty sectors, instead of active, in post proof
In actors v2, active doesn't include "new" (unproven) sectors.
2020-09-30 17:35:14 -07:00
Łukasz Magiera
90be8d801b
Merge branch 'master' into post-trace 2020-09-30 10:03:54 +02:00
Łukasz Magiera
7a3a2f8db9
Merge pull request from filecoin-project/feat/dont-recompute-PoST-redux
Dont recompute post on revert <<redux>>
2020-09-30 09:18:46 +02:00
Łukasz Magiera
ef28ebb14a Ignition upgrades, much excite. 2020-09-25 23:55:37 -07:00
Dirk McCormick
80a7ed8116 refactor: use abstract types instead of specs-actors 2020-09-25 15:54:27 +02:00
zgfzgf
13b38815c2 add trace wdpost 2020-09-25 20:31:07 +08:00
Dirk McCormick
7eb9bec13f feat: dont recompute post on submit redux 2020-09-24 14:38:33 +02:00
Steven Allen
63f026f7c3 migrate some more imports 2020-09-21 15:24:45 -07:00
Steven Allen
4cf0c105eb optimize sector loading
And avoid exposing "arrays" via the miner abstraction. We may change these
structures later.
2020-09-21 12:12:08 -07:00
Łukasz Magiera
a307e4593a wdpost: Fix TestWDPostDoPost 2020-09-21 10:06:19 +02:00
Steven Allen
35bce5a5c6 revert post changes
1. Calling the specific partition/deadline APIs is faster.
2. It's _much_ easier to test this way.
2020-09-18 15:40:49 -07:00
Steven Allen
1bf3b4989d rename imports to match actors code
`sed -i 's/\bv0\(\w\)\(\w*\)/\L\1\E\20/g' **/*.go`
2020-09-18 14:59:27 -07:00
Łukasz Magiera
ccd0a67b01 wip fixing wdpost tests 2020-09-18 20:34:23 +02:00
Łukasz Magiera
70faa36b7f Merge remote-tracking branch 'origin/master' into refactor/net-upgrade 2020-09-18 19:29:06 +02:00
Łukasz Magiera
26bbd9fb9d wdpost: append is hard 2020-09-18 03:00:31 +02:00
Łukasz Magiera
8419dda84b wdpost: Only add proven partitions to message params 2020-09-18 02:39:08 +02:00
Łukasz Magiera
6eda53565f Most tests passing 2020-09-17 17:30:24 +02:00
Aayush Rajasekaran
31ff5230bb Get State API almost working 2020-09-17 05:05:32 -04:00
hannahhoward
691bd9f442 feat(markets): complete markets conversion
complete markets conversion to using chain/actors types, also replacing DealProposal/DealState
interfaces with structs
2020-09-17 00:43:14 -07:00
Aayush Rajasekaran
b5ba7a0fad Miner migration 2020-09-17 02:21:16 -04:00
jennijuju
5c69249ba3 Use window post for window PoSt related log messages. 2020-09-15 21:22:29 -04:00
Steven Allen
c64f983900 migrate StateMinserSectorCount 2020-09-15 12:09:39 -07:00
Łukasz Magiera
4dabab5ce6 state manager progress 2020-09-15 13:04:45 +02:00
Dirk McCormick
db998e9c0f fix: storage - move rand generation after proofs 2020-09-15 09:52:55 +02:00
Dirk McCormick
f0f15f899c feat: split window PoST messages into batches 2020-09-15 09:43:18 +02:00
Steven Allen
36f920bcd7 progress 2020-09-14 21:55:49 -07:00
Steven Allen
02dcb5e182 Merge branch 'master' into refactor/net-upgrade 2020-09-14 14:53:57 -07:00
Steven Allen
2088335b9d rename to simplify merge 2020-09-14 14:40:52 -07:00
Steven Allen
9804310cc8 Update specs-actors to 0.9.10
Moves the Keyer type (and helpers) into go-state-types.
2020-09-14 14:32:57 -07:00
Łukasz Magiera
8a05fe86d6 wdpost: Skip sectors correctly 2020-09-14 20:33:54 +02:00
Łukasz Magiera
06ec571c20 Merge remote-tracking branch 'origin/master' into refactor/net-upgrade 2020-09-14 15:09:52 +02:00
Raúl Kripalani
6d29d75724 Merge branch 'master' into inmem-journal 2020-09-14 12:17:45 +01:00
Steven Allen
d3594835c4 [WIP] Network upgrade support
This patch starts adding support for network upgrades.

* It adds an actors abstraction layer for loading abstract (cross-version) actors.
* It starts switching over to a shared deadline type.
* It adds an abstraction for ADTs (hamt/amt).
* It removes the callback-based API in the StateManager (difficult to abstract
across actor versions).
* It _does not_ actually add support for actors v2. We can do that in a followup
patch but that should be relatively easy.

This patch is heavily WIP and does not compile. Feel free to push changes
directly to this branch.

Notes:

* State tree access now needs a network version, because the HAMT type will change.
* I haven't figured out a nice way to abstract over changes to the _message_
types. However, many of them will be type aliased to actors v0 in actors v2 so
we can likely continue using the v0 versions (or use the v2 versions
everywhere). I've been renaming imports to `v0*` to make it clear that we're
importing types from a _specific_ actors version.

TODO:

* Consider merging incremental improvements? We'd have to get this compiling
again first but we could merge in the new abstractions, and slowly switch over.
* Finish migrating to the new abstractions.
* Remove all actor state types from the public API. See `miner.State.Info()` for
the planned approach here.
* Fix the tests. This is likely going to be a massive pain.
2020-09-11 20:16:29 -07:00
Łukasz Magiera
d0175392c0 Merge remote-tracking branch 'origin/master' into feat/window-post-faulty-sectors 2020-09-11 10:47:39 +02:00
Aayush Rajasekaran
774e068436 Update to specs-actors v0.9.8 2020-09-10 17:41:55 -04:00
Łukasz Magiera
5e7737f55d wdpost: Handle skipped sectors correctly 2020-09-10 02:59:37 +02:00
whyrusleeping
7a6ceebb34 windowed post generation now returns faulty sectors 2020-09-09 14:00:15 -07:00
Steven Allen
afa913f2ed Remove hard-coded late-fee in window PoSt
This is no longer necessary, and doesn't really serve any purpose.
2020-09-09 11:34:55 -07:00
Aayush Rajasekaran
d678fe4bfa Fix tests 2020-09-07 15:48:42 -04:00
Aayush Rajasekaran
39755a294a Update to specs v0.9.6 2020-09-07 15:48:41 -04:00
Raúl Kripalani
3206f92063 Merge branch 'master' into inmem-journal 2020-09-02 19:50:52 +01:00
Raúl Kripalani
f046af337a split wdpost event into finer-grained ones. 2020-09-02 19:45:25 +01:00
Raúl Kripalani
ac152abc75 shuffle code to journal wdpost events. 2020-09-02 19:15:25 +01:00
Łukasz Magiera
7806a9885a storagefsm: Record publish deals message CID 2020-08-27 17:50:50 +02:00
Steven Allen
0155a31d1f Fix PoSt with bad sectors
"skipped" sectors must be replaced with a substitute "good" sector, or the
entire partition must be skipped. They should not just be omitted.

This patch also fixes the test to verify the _entire_ proof instead of just
verifying that the proof includes the correct sectors.
2020-08-26 09:56:51 -07:00
Raúl Kripalani
efdfd3ee3e Merge branch 'master' into inmem-journal 2020-08-26 16:38:23 +01:00
Raúl Kripalani
4e1ef09751 make journal a global var. 2020-08-26 16:09:37 +01:00
Łukasz Magiera
a236957538
Merge pull request from filecoin-project/asr/marked-for-upgrade
Indicate whether a sector has been marked for upgrade
2020-08-22 21:48:45 +02:00
Aayush Rajasekaran
984e52acbc Indicate whether a sector has been marked for upgrade 2020-08-21 15:53:39 -04:00
Steven Allen
5733c71c50 Lint everything
We were ignoring quite a few error cases, and had one case where we weren't
actually updating state where we wanted to. Unfortunately, if the linter doesn't
pass, nobody has any reason to actually check lint failures in CI.

There are three remaining XXXs marked in the code for lint.
2020-08-20 20:46:36 -07:00