Github -> Issues -> Here -> Search -> Gateway -> ipfs.io -> dweb.link

Help IPFS
#24

Hrm, it really depends on your OS and which way you installed IPFS initially. Iā€™m not sure you can use ipfs-update in every case (Iā€™ve never used it myself). If you are not installing on a server, the simplest way is to use IPFS Desktop.

#25

Iā€™m running go-ipfs with Ubuntu LTS latest and let it install in the default location. I only have Desktop installed on my old Win7 PC and I do like its update feature. Wish go-ipfs was as easy to update.

Iā€™m going to try running ipfs-update from the .ipfs directory. That produced the same error.

#26

the simplest way to replace it is to shut down the currently running instance (looks like that already happened), download the appropriate binary from:

check where your current install is with:

which ipfs

remove and replace the binary from that location (you may have to use sudo), run:

ipfs --version

to verify itā€™s installed properly, and then launch the daemon the usual way you do it.

P.S. the first time you run 0.12, it will migrate your repo, which may take a while if you have a lot in it.

P.P.S Here is the official version of what I just explained:

#27

That worked. Many thanks. Everything now look okay?

#28

Looks good from here, and running 0.12! :slight_smile:

#29

Appreciate all your help. Time for a cool beverage and a movie :sunglasses:

#30

Are there any ports other than 4001 which my firewall should allow from outside to optimize go-ipfs for access to the websites I publish on IPFS? That includes all the ones listed with a publishing program link at
ipns://k2k4r8n098cwalcc7rdntd19nsjyzh6rku1hvgsmkjzvnw582mncc4b4/IPFS_WebsitesDirectory.html

#31

Your node appears to use a random port every time you restart it. It is probably using UPnP to acquire a port instead of using 4001. Check your config at Swarm.DisableNatPortMap (I have mine set to true, so it doesnā€™t do that). If thatā€™s the case, you actually donā€™t need to set up a static port forwarding (thatā€™s what UPnP does for you). If youā€™d rather use static port forwarding, set that to true and port map UDP 4001 to 192.168.33.135:4001 and TCP 4001 to 192.168.33.135:4001, then make sure you are still reachable. Also, make sure that Reprovider.Strategy is set to ā€œallā€.

#32

Swarm.DisableNatPortMap was false and now is true
Reprovider.Strategy was already set to ā€œallā€
Static IP .135 was already set on the node PC
Restarted the daemon after the config changes.
Port forwarding was already TCP and now also UDP.
Will reboot the node PC as soon as the current publish completes.

Many many thanks

#33

The sites already seem to be responding much better, even without the node reboot.

#34

I donā€™t see your node anymore, something is wrong with its advertising. Type ā€œipfs idā€ at the terminal and see which addresses itā€™s currently advertising, and make sure it includes both public addresses.

#35

It is now advertising some addresses, but none of them are working:

> ipfs id 12D3KooWHWen2dsu4QWJE7srfSFduVPKTnGQ8jbx5MtG7iQc7YWD
Error: failed to dial 12D3KooWHWen2dsu4QWJE7srfSFduVPKTnGQ8jbx5MtG7iQc7YWD:
  * [/ip4/104.156.24.117/tcp/54136] dial tcp4 104.156.24.117:54136: connect: connection refused
  * [/ip6/64:ff9b::689c:1875/tcp/54136] dial tcp6 [::]:4001->[64:ff9b::689c:1875]:54136: i/o timeout
  * [/ip4/104.156.24.117/udp/54136/quic] timeout: no recent network activity
  * [/ip4/104.156.24.117/udp/4001/quic] timeout: no recent network activity
> nc 104.156.24.117 4001
>

I even tried to netcat into TCP 4001, and that doesnā€™t respond either. So, Your node is once again unreachable and needs some tlcā€¦

#36

It looks like the internet connection may have dropped for a few minutes a little bit ago.
ipfs id on this end now shows all the TCP and UDC connections without any errors or warnings

#37

Still getting the same results, I donā€™t think either of your 2 port forwards are actually working at all. If you canā€™t get them to work, you should re-enable UPnP, which will make your node reachable again.

#38

For some reason, after the internet interruption the peer count never recovered. Just restarted the daemon. Peer count is recovering and the sites on IPFS seem to be responding again.

#39

But what about your inbound count (I still canā€™t connect)?

#40

Inbound was good before my nap with a large number. Internet drop out seems to have knocked the daemon down since then with only a couple remaining inbound. Having an ice storm today and the internet is up and down. Just restarted the daemon again.

#41

Yay, it just started to work properly:

> ipfs dht findpeer 12D3KooWHWen2dsu4QWJE7srfSFduVPKTnGQ8jbx5MtG7iQc7YWD
/ip6/::1/udp/4001/quic
/ip4/192.168.33.135/tcp/4001
/ip6/64:ff9b::689c:1875/tcp/4001
/ip6/::1/tcp/4001
/ip4/127.0.0.1/udp/4001/quic
/ip4/127.0.0.1/tcp/4001
/ip4/192.168.33.135/udp/4001/quic
/ip4/104.156.24.117/tcp/4001
/ip4/104.156.24.117/udp/4001/quic

The only thing is that it advertises an IPv6 address, but doesnā€™t answer on it (it might be an error).

P.S. You can count the IPv6 inbounds with the following command:

> ipfs swarm peers --direction | grep '^/ip6/.\+\sinbound$' | wc -l
      64
#42

That command without the 64 at the end returns 0 inbound ipv6
Not surprised. Last I checked my ISP wasnā€™t yet supporting ipv6

#43

The 64 was my result when I ran it on my own node, sorry. I donā€™t understand why your node is advertising an IPv6 address then.

An easy way to get rid of it is to remove the 2 ip6 lines from Addresses.Swarm (which is fine, since your node doesnā€™t use IPv6).

Also, your node has begun to advertise some more ports too, something weird is going on:

> ipfs dht findpeer 12D3KooWHWen2dsu4QWJE7srfSFduVPKTnGQ8jbx5MtG7iQc7YWD
/ip4/104.156.24.117/tcp/4001
/ip4/104.156.24.117/udp/1026/quic
/ip6/::1/tcp/4001
/ip4/127.0.0.1/udp/4001/quic
/ip6/64:ff9b::689c:1875/tcp/4001
/ip6/::1/udp/4001/quic
/ip4/104.156.24.117/udp/4001/quic
/ip4/104.156.24.117/tcp/1024
/ip4/192.168.33.135/tcp/4001
/ip4/127.0.0.1/tcp/4001
/ip4/192.168.33.135/udp/4001/quic