Signal-desktop: Recent upgrade no "Disconnected"

[massiveradiation]

Launch app, Mac sierra 10.12.6. Says disconnected. Wifi is present and active. No changes to the system. Attempting reconnect, but then nothing.

[Fancellu2]

Signal seems down. Signal doesn't have a status page.

This is not an app issue.

[GuardianMajor]

On windows here and I was about to post the same thing but saw this. Since the update, completely unable to use it and stuck in a constant retrying in 60 seconds loop. Even though there is nothing wrong with the connection.

But as @fancellu2 suggested, it seems to be a service issue since android can't send or receive anything either.

[Hoorah]

Yep, I'm experiencing the same thing -- "Disconnected .. Attempting reconnect in 60 seconds" -- but Signal never reconnects. Quitting Signal and opening it again does nothing to clear up the problem. This occurred right after I accepted Signal's invitation to upgrade my Signal Desktop app to
v1.12.0.

[cfoster]

Mac OS: Signal has been down for a while. It was down on a previous version, I got notified of a new version (1.12.0), downloaded/restarted, still doesn't connect properly.

[cleverlemming]

I upgraded today to v1.12.0 for OSX on a Mac Retina OSX version 10.13.4. No successful subsequent connection/constant "Disconnected. Attempting reconnect..." messages. Messages will not send.

[joanmarcriera]

Same here,I was about to create a new bug but searched a bit.

macOs high sierra with signal app 1.12.0

[keeepinitgansta]

I remember that amazon was threatening to shutdown Signal's AWS server access.... I hope amazon didn't finally make good on their threat. :-(

https://signal.org/blog/looking-back-on-the-front/

AWS seems to have some issues with servers in the east coast..

https://status.aws.amazon.com/

[ae-g-i-s]

DNS for textsecure-service.whispersystems.org points to 127.0.0.1, so it seems that there's either a misconfiguration or something more sinister altogether (TM).

8.8.4.4, 8.8.8.8 and various online resolvers all return this result.
textsecure-service-staging.whispersystems.org still returns what seems to be a sane result (untested), the CDN servers still point to Cloudfront. Only the "main" server points to localhost.

The Amazon threat @keeepinitgansta is referring to can be found at https://signal.org/blog/looking-back-on-the-front/, FWIW. Not sure it is the reason for this outage, but it's possible.

Reddit discussion: https://www.reddit.com/r/signal/comments/8nmhfk/signal_down/ (no "me too" pls, might get infos faster though).

[keeepinitgansta]

You're right @ae-g-i-s . DNS records are pointing to localhost (127.0.0.1) for all DNS servers (public / private). They need to update this pronto!

[gburtini]

Confirming that IPv4 textsecure-service is broken.

~ $ ping textsecure-service.whispersystems.org
PING textsecure-service.whispersystems.org (127.0.0.1) 56(84) bytes of data.
64 bytes from localhost (127.0.0.1): icmp_seq=1 ttl=64 time=0.044 ms
64 bytes from localhost (127.0.0.1): icmp_seq=2 ttl=64 time=0.036 ms
^C
--- textsecure-service.whispersystems.org ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1030ms
rtt min/avg/max/mdev = 0.036/0.040/0.044/0.004 ms
~ $ dig A textsecure-service.whispersystems.org

; <<>> DiG 9.12.1 <<>> A textsecure-service.whispersystems.org
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 6938
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;textsecure-service.whispersystems.org. IN A

;; ANSWER SECTION:
textsecure-service.whispersystems.org. 59 IN A  127.0.0.1

;; Query time: 17 msec
;; SERVER: 192.168.0.1#53(192.168.0.1)
;; WHEN: Thu May 31 16:16:56 PDT 2018
;; MSG SIZE  rcvd: 82

~ $ dig AAAA textsecure-service.whispersystems.org

; <<>> DiG 9.12.1 <<>> AAAA textsecure-service.whispersystems.org
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 61036
;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;textsecure-service.whispersystems.org. IN AAAA

;; AUTHORITY SECTION:
whispersystems.org. 531 IN  SOA ns-1008.awsdns-62.net. awsdns-hostmaster.amazon.com. 1 7200 900 1209600 86400

;; Query time: 15 msec
;; SERVER: 192.168.0.1#53(192.168.0.1)
;; WHEN: Thu May 31 16:16:59 PDT 2018
;; MSG SIZE  rcvd: 151

[keeepinitgansta]

I think this is either a Stale DNS record or an Amazon AWS server issue.

[cfoster]

The name-servers for the domain "whispersystems.org" which will ultimately control the subdomains, including "textsecure-service" are:

Name Server: NS-1008.AWSDNS-62.NET
Name Server: NS-43.AWSDNS-05.COM
Name Server: NS-1381.AWSDNS-44.ORG
Name Server: NS-1752.AWSDNS-27.CO.UK

Given that and the fact they were given some kind of warning about 30 days ago from AWS, my money would be on AWS changing the IP entries here / and or - blocking service altogether. I may be wrong.

[cfoster]

Whether AWS has blocked service or not. It would be good to at least get some sort of an update on the Twitter feed or website as to what the real status is instead of letting us all try to work out what the real issue is here.

[popey]

Looks like it's slowly coming back. Managed to get into the app and re-associate with my phone. But I see no contacts or messages. Guess I'll wait and look again in the morning :)

[cfoster]

It's working again.
You really need to make a statement.
The fact it went down for so long, DNS changing/faulty, pushing a new version.
You haven't been taken over by some 3rd party wanting to snoop have you?

[scottnonnenberg-signal]

https://twitter.com/signalapp/status/1002336532903022592

[cfoster]

Why would you set IP entries to localhost and release a new piece of software for a power outage?

There are plenty of other providers of open source end-to-end encryption apps. If you can't properly give an answer to this you're gonna' lose everyone.

[scottnonnenberg-signal]

Hey everyone, just wanted to provide some clarification beyond that tweet. Today's Signal Desktop app update was unrelated to the outage; it was released before the outage occurred. The DNS entries were changed (by us) as a way of shedding some traffic during recovery. We apologize for the inconvenience.