Closed FreddieRidell closed 2 years ago
Which version of hypercore are you using? On Fri, 18 Jan 2019 at 13.29, Freddie Ridell notifications@github.com wrote:
I've recently started seeing a lot of "No proof available for this index" errors being created while trying to replicate a hyperdb https://github.com/mafintosh/hyperdb using hyperswarm https://github.com/hyperswarm/network.
The error is being created here: https://github.com/mafintosh/hypercore/blob/547a68ce2a32a9be58a2c85327465191b5a41c70/index.js#L402
I'm not enteirly sure what's causing this, as I don't believe I'm doing anything that could be damaging the integrity of the underlying hypercores.
Do you have any advice on what causes this error, and how it could be prevented/mitigated/recovered from?
PS. I really love all the hyper-suite of packages, Dat, Beaker and the work being done here! Thanks so much for your contibutions and I'm looking forward to multi-writer Dat π
β You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/mafintosh/hypercore/issues/191, or mute the thread https://github.com/notifications/unsubscribe-auth/AAW_VdgSfTpfbNzupvJztcVnoTFJCapiks5vEb4WgaJpZM4aH0WT .
6.22.4
Oh thatβs weird then! Do you have an example or something to help me debug this? That should def not happen
Actually just realised that the 6.22.4 has to run on the receiving peer for the bugfix to work. Do you know if that is the case here? Still a test case or example would be great to fix this otherwise :)
This has unfortunately been quite an intermittent bug, probably caused by bad socket connections between replicating hyperdbs. If I can reproduce the issue I'll tar
up my data and post it here?
Yea that would be great, thanks On Mon, 21 Jan 2019 at 19.14, Freddie Ridell notifications@github.com wrote:
This has unfortunately been quite an intermittent bug, probably caused by bad socket connections between replicating hyperdbs. If I can reproduce the issue I'll tar up my data and post it here?
β You are receiving this because you commented.
Reply to this email directly, view it on GitHub https://github.com/mafintosh/hypercore/issues/191#issuecomment-456160115, or mute the thread https://github.com/notifications/unsubscribe-auth/AAW_VTSurHTg36T8oY_JX9N9jmI654Slks5vFgN_gaJpZM4aH0WT .
Ok, I've got a reproduction. I have a hyperdb replicated between two peers:
I replicated it using https://github.com/mafintosh/discovery-swarm and now when I try to replicate them again my laptop fails with the following error:
Error: No proof available for this index
at Feed.proof (/home/freddie/.config/yarn/global/node_modules/hypercore/index.js:402:25)
at Peer.onrequest (/home/freddie/.config/yarn/global/node_modules/hypercore/lib/replicate.js:130:13)
at Feed._emit (/home/freddie/.config/yarn/global/node_modules/hypercore-protocol/feed.js:182:32)
at Feed._onmessage (/home/freddie/.config/yarn/global/node_modules/hypercore-protocol/feed.js:162:10)
at Protocol._onmessage (/home/freddie/.config/yarn/global/node_modules/hypercore-protocol/index.js:344:6)
at Protocol._parseMessage (/home/freddie/.config/yarn/global/node_modules/hypercore-protocol/index.js:393:10)
at Protocol._parse (/home/freddie/.config/yarn/global/node_modules/hypercore-protocol/index.js:358:37)
at Protocol._write (/home/freddie/.config/yarn/global/node_modules/hypercore-protocol/index.js:265:8)
at doWrite (/home/freddie/.config/yarn/global/node_modules/readable-stream/lib/_stream_writable.js:428:64)
at writeOrBuffer (/home/freddie/.config/yarn/global/node_modules/readable-stream/lib/_stream_writable.js:417:5)
Emitted 'error' event at:
at errorOrDestroy (internal/streams/destroy.js:98:12)
at Protocol.onerror (_stream_readable.js:717:7)
at Protocol.emit (events.js:197:13)
at Protocol.destroy (/home/freddie/.config/yarn/global/node_modules/hypercore-protocol/index.js:227:17)
at Feed.destroy (/home/freddie/.config/yarn/global/node_modules/hypercore-protocol/feed.js:92:15)
at Peer.destroy (/home/freddie/.config/yarn/global/node_modules/hypercore/lib/replicate.js:400:15)
at onproof (/home/freddie/.config/yarn/global/node_modules/hypercore/lib/replicate.js:133:26)
at Feed.proof (/home/freddie/.config/yarn/global/node_modules/hypercore/index.js:402:22)
at Peer.onrequest (/home/freddie/.config/yarn/global/node_modules/hypercore/lib/replicate.js:130:13)
at Feed._emit (/home/freddie/.config/yarn/global/node_modules/hypercore-protocol/feed.js:182:32)
the files from my laptop: https://send.firefox.com/download/6113755815/#tGDBtTpykxugmFkmCcSQ_A the files from my server: https://send.firefox.com/download/6cc51092a4/#QH2IycW9d1i_eoPc5EbfUA
They contain semi-personal info, so these links expire in 24 hours, I can re-upload if needed :)
Thanks for your help
@FreddieRidell thanks, i downloaded the files. will try and repro over the weekend
I am also running into this error when i move the hyperdb from one machine to another behind different routers on different networks. This actually ended up being a problem with me running two different versions of hyperdb that were not compatible.
Closing cause old and lots have changed
I've recently started seeing a lot of
"No proof available for this index"
errors being created while trying to replicate a hyperdb using hyperswarm.The error is being created here: https://github.com/mafintosh/hypercore/blob/547a68ce2a32a9be58a2c85327465191b5a41c70/index.js#L402
I'm not enteirly sure what's causing this, as I don't believe I'm doing anything that could be damaging the integrity of the underlying hypercores.
Do you have any advice on what causes this error, and how it could be prevented/mitigated/recovered from?