Release 0.11.0

[hsanjuan]

• Default allocations filter should be all #762 Fix /allocations endpoint when no filter used

[hsanjuan]

• Update deps #769 deps
• [Proxy] Intercept pin/update #732 Proxy: hijack pin/update #768 proxy hijack
• Types: make UserAllocations []peer.ID instead of string #767 UserAllocs peer ID

[hsanjuan]

• RPC: Give each component a different RPC Service #773: RPC services for each component

[hsanjuan]

• Feat: RPC Authorization #775 RPC Authorization #710 Authorization for RPC #666 RPC: Give each component a different RPC Service #773 RPC Auth

[hsanjuan]

• Separate identity and configuration #760 Separate Identity and Configuration #766 service: multiple fixes around init and identities #780 Separate identity and configuration

[hsanjuan]

• Run tests with -race #340 Fix #340: Run tests with -race #502 Tests: add failfast, bump to go1.12, adjust codecov #779 Improve race conditions in tests et al

[hsanjuan]

• Repinning should not rely on Leader() #413 Change the underlying data structure of a metrics.Window #713 Monitoring and re-allocation revamp: Accrual failure detection #714 accrual monitoring and leaderless repinning
• Monitor: alerts never stop #812 Monitor: fake alerts #813 alerting for peers stops after one alert #814 fix how accrual fd treats ttls #815 related

[hsanjuan]

• Update go-ds-crdt #789 go-ds-crdt update
• expose badger loading mode conf #776 expose badger conf Daemon using CRDTs fails to start on Raspberry Pi #771

[hsanjuan]

• Both consensus are bleeding peers #787 Feat: introduce a ConnectionManager for the libp2p host #791 Fix #787: Connectivity fixes #792 improve connection handling (changes config)

[hsanjuan]

#795 Deps updates

[hsanjuan]

#685 new "crdt" consensus component (can't believe i forgot to put it here)

#804 tracing for hooks etc

#811 fixes: #787 #798 #805

#816 #820

#856 make consensus choice an explicit requirement in service.

[lanzafame]

#806 Fix: api: replace two regexes with strings.HasPrefix()

[hsanjuan]

#808 #809 Fix base32

[lanzafame]

#807 add IPFSPinStatusFromString benchmark

[hsanjuan]

#818 : Fix importing peers with /dnsaddr/

[hsanjuan]

#822 #823 Fix badger dep

[hsanjuan]

#765 #781 Remove gx from makefile

[kishansagathiya]

#497 #738 Pin timeouts should start from the last block

[eLBhogi]

@hsanjuan
(1) Can you please detail timeline to finalise this release, and what is the procedure to upgrade the cluster to 0.11 + upgrade go-ipfs to 0.4.21 (zero-downtime, zero state-loss method)

(2) Can we upgrade just go-ipfs to 0.4.21 without upgrading ipfs-cluster to 0.11? If so, what is the safe way to do this without losing downtime or state.

Thanks

[hsanjuan]

(1) Can you please detail timeline to finalise this release, and what is the procedure to upgrade the cluster to 0.11 + upgrade go-ipfs to 0.4.21 (zero-downtime, zero state-loss method)

Hi. I want to land a couple of features plus have the documentation re-write finished before releasing. I would like to ship this release in 4 weeks at most. Docs are the main blocker and I think most issues are resolved in the latest RC.

The upgrade procedure

There's some info here: https://cluster.ipfs.io/documentation/upgrades/ . Usually, shut all peers down, upgrade, and start them again. This has a few seconds of downtime though. I'm not sure on the details of your system though to see if it would be possible to approach 0 downtime in some way. We'll introduce some improvements on this front for next releases.

(2) Can we upgrade just go-ipfs to 0.4.21 without upgrading ipfs-cluster to 0.11? If so, what is the safe way to do this without losing downtime or state.

Yes, just upgrade go-ipfs and don't touch the cluster peers.

[hsanjuan]

#831 auto-recover (config change)

[hsanjuan]

#793 #832 disable unpinning (config change)

[hsanjuan]

#833 fix flaky add test

[hsanjuan]

#828 #830 badger build fix

[hsanjuan]

#837 #778 fix config name not matching env vars.

[hsanjuan]

#452 #836 Do not load API components if removed from the config

[hsanjuan]

#448 #737 Load balancing client for rest API client.

[hsanjuan]

#843 Improve Cluster.Pin/Unpin method signatures (changes Go API, changes Go Rest Client, Changes API responses for pin/unpin to match with pinpath/unpinpath).

[hsanjuan]

#835 #839 init --peers #870

[hsanjuan]

#853 #855 Fix emitter panic on API shutdown
#742 #854 Fix API not returning 404 on delete of unexisting pins.

[hsanjuan]

#840 #846 Remove Raft peers from peerstore on removal

[hsanjuan]

#857 #834 auto-trust bootstrappers

[hsanjuan]

#648 #860 fix: peers are removed from peerstore when they're down. (dependency update)

[hsanjuan]

• Update ipfs-cluster to make use of hRaft LeaderObservation type  #211 Make use of hRaft LeaderObservation type #858 LeaderObs type cleanup

[hsanjuan]

#856 #862 make --consensus mandatory

[hsanjuan]

#863 more test improvements

[hsanjuan]

#868 remote configuration
#864 #803 follower mode config option
#867 trust all by default

[hsanjuan]

#870 fix sharness (add to init --peers)

[hsanjuan]

#872 enable goproxy on dockerfiles
#871 #852 Improve error handling on add
#869 #732 Proxy/update done right
#868 Support remote configuration for daemon
#874 Select consensus on init (configuration restriction)
#875 Remove pin_method
#882 : docker bump to crdts PR

[hsanjuan]

#876 string flag rather than stringslice

[hsanjuan]

#657 #879 make sure API errors are JSON
#885 update go-ds-crdt
#886 reading peerstore file panics on empty lines
#887 fixes: pin update set with undef cid, increase status hostname spaces, pintracker spamps the logs

[hsanjuan]

#899 Warn about trustall
#900 mdns new

[hsanjuan]

#906 fix crdt mass import, avoid compacting badger
#905 fix trust all remained enabled

[hsanjuan]

#431 #877 /p2p/ multiaddresses
#761 #890 user allocations for add
#377 #901 pin queue maptracker
#797 #892 use latest sharness

[hsanjuan]

#670 #891 support metadata in ipfs-cluster-ctl
#327 #878 Sort addresses and peer lists

[hsanjuan]

#888 #912 Fix allocations handling
#913 Consolidate parsing of pin options

[hsanjuan]

#574 #894 logging for restapi

[hsanjuan]

#755 #909 improve replication factor 0 explanation

[hsanjuan]

#918 more crdt update

[0zAND1z]

Just curious. When is v0.11.0 planned to show up on dist.ipfs.io?

[hsanjuan]

@0zAND1z it is there (all versions). I made a mistake but I'll fix that now