Has anyone elsa been having problems with the Icelandic exit nodes for the past couple of days? I have had trouble with AUTH failures on linux and intermittent drops in the connection usually in 30-40 minutes or so. I am wondering if this is a problem with the nodes or if it is my new router. I guess all my emails and messages to the cryptostorm team go to the spam folder

I haven't personally experienced any flakiness on iceland, but I've passed this on to team here to investigate

@OP

I've had this instability recently on the German nodes, both pre- and post-Heartbleed instances. I've also had AUTH failures on the Iceland node, but as I'm writing this it has authenticated successfully.

via mobile I've had connctions dropping on bruno, onyx, frankfurt, and fenrir this week.
Not sure if it's my 3G connection, as Linux seems to be ok

I am getting both soft auth failures with networkmanager on linux and a curious thing...after some time when I have finally managed to connect the connection chokes, that is, my connection wont work and I get a timeout error....nedless to say...this is driving me a little crazy

I'm getting knocked off every node I connect to within 5-10 minutes on each connect. I've seen a few token auth failures also but was able to reconnect ...for another 5 minutes...

I can verify I am being either disconnected or staying connected with no access to the Internet. All nodes show this behavior

We're seeing a number of people with this issue, and unfortunately we do not have a fix just yet. We're noticing a high number of people having issues are running some flavour of Linux or using viscosity on OSX. Is that the case with everyone here?

Linux here

@cryptostorm_ops

When I had issues, yes it was on Linux.

@cryptostorm_ops

I have the same problems as the others. It has been for some days but i am using Windows 7 64 bit.

OK, we've seen enough of these complaints that we have to assume that we're missing something. We've noticed that a lot/most of the issues are via Linux, thus we're wondering if some of you with intermittent issues can email your .conf files to [email protected] so we can see if there's some setting that's causing issues.

Thanks, and apologies for the issues!
G

Quark5 wrote:@cryptostorm_ops

I have the same problems as the others. It has been for some days but i am using Windows 7 64 bit.

Oh, just caught up to see this Win7 in there. Hmmm. Maybe it's not Linux related. Are you using our widget or OpenVPN directly? If it's OpenVPN can you send us your conf file as well, please?

Thank you!

@Graze

I've just sent my config (Montreal) which I was using to test Bruno (174.142.78.196). One thing I've noticed: every other node that I've used shows the exit IP and connection IP as being the same. However, I connect to Bruno on 174.142.78.196, yet I exit on 174.142.78.193.

Is this correct?

@Graze - I have a dual boot system of Linux Mint/Win7 but the problems I have experienced yesterday were only in my Win7 boot, I hadn't tried my Linux boot. I'm using the widget when I'm in Win7.

@Graze

I am using the widget (1.10). I usually use the german and the french nodes and the problem is on both of them.

Also experiencing issues, Linux Mint 14.04 - exactly as described, works for 10ish minutes then drops, have to wait a few minutes before it will reconnect again. Using OpenVPN 2.3.2

So is anything at all coming up. I have sent my config to the cryptostorm support, but as it is a vanilla configuration my suspicion is that the fault lies with the servers. Are there any other exit nodes in Iceland that I can try?

We are still looking into it - it's what one might call "a bit of a puzzler" in that it was stable before and seems to no longer be, despite nothing of substance changing server side - we've had lots of discussion on it, which led to this little bit of sunshine which basically says, "it's better to drop the connection, than to connect insecurely."

Obviously our goal is to actually connect both solidly AND securely. We continue to look into it.

@gbj

I have to agree with you. I've just checked through the client-side configs for Onyx (a post-Heartbleed instance, exclusively) and Cantus (pre-Heartbleed instance). The parameters are identical apart from the connection addresses, so the issue must indeed be server-side.

@ALL_USERS

If ever there was a time to deploy/double check the operation of your firewall, this is it.

same disconnect problems on bruno / chilli and emerald
win XP and win7 using openvpn raw

chilli and emerald started disconnecting several days ago,
then bruno started disconnecting yesterday

Should be better now(?)

DF made some server mods, and we think that may have helped.

Any confirmation from you guys??

Thanks!
G

Yes. The connection is much better now. I have not seen any AUTH errors in my logs for the past hours. My connection still drops without any warning or reason, I cannot find anything in my logs to explain it. I usually get 30 minutes to an hour connected until the connection times out.

I hate to bring the bad news, I was on bruno yesterday with no interruptions but today on onyx I've lost connection. Each time that happens I have token auth failed 2-3 times when I try to reconnect. I get that error whether the connection drops or even if I disconnect the widget manually. For now... back to bruno and see how that goes again

I've had no log outs in the last ~ 6 hours via linux and bruno
Real test is tomorrow on mobile

Edit: except now... although it could have been my dodgy internet coonection?

I still have the drops in the connection i used Germany (cantus). I see no improvment.

@ops

Connected to Onyx today, connection lasted 25 mins then dropped.

I guess I spoke too soon, the situation has somewhat improved on fenrir, but auth errors and timeouts are still happening often, every 30 minutes or so it seems

@ops

Tested on Cantus, same again. 25 mins uptime.

Hi, I am not sure if this is the right area to post in but here goes...

For the past week or so I have had things like: Invalid token (But I have paid for a years subscription), this tends to resolve itself after a while and I can then start my VPN.

The main problem is that when I have cryptostorm on I have started to lose internet connection so frequently that I can't even download a small file. When this happens my internet bugs out and I can't open pages and if I am playing games I tend to lose everything. My torrent program does eventually start up again but after a long time and I am not completely sure that when this happens I am even registering as not being in Sweden....

Why has this started happening? I have not changed anything on my computer and losing internet connection only happens when I have Cryptostorm on.

Help?

@OP

Yes there are issues which the team and the community are working to iron out. Please give more details of your operating system etc, but please do so in this thread, so we can keep everything in one place. Many thanks.

@ops

Quick update. Tested on Fenrir, managed to get 2h20mins out of it before it dropped.

Add me to the list I am in Sweden and have been dropping internet connection a lot. So much that I can barely download anything without disconnecting. When this happens my torrent program will re connect but after a long time- however my internet bugs out and I can't open pages. I have to disconnect from Cryptostorm and start over.
I also have had a message that my token is invalid a few times, but after a few goes it usually connects.

I am using Windows 7 64 bit operating system.

@kiwifruit

Are you using the CS widget or the Windows OpenVPN client?

Hi,

my router with asuswrt/merlin software has disconects every few minutes on every canada and the iceland server (after a few weeks working with no problem).

@parityboy I am pretty sure I used the widget back in Feb...

OK, so parityboy and fermi worked with us tonight on this issue via IRC. We think we have a fix, but again (and ya, this is a different fix) I need confirmation: Are you folks getting better connections now? What I mean is that some people were experiencing regular disconnections every twenty minutes or so - I am hoping that those are no longer happening.

Thanks again so much to parityboy, fermi, tls, and all the others who were working with us in chat on this issue!

So far, nothing disastrous to report... online and uninterrupted for an hour and a bit!

Can I ask what config files you guys are using? I am still being dropped from fenrir on my linux box :/

Can I ask what config files you guys are using? I am still being dropped from fenrir on my linux box :/

Hiya man, here's a link to config files.

My Android connection was getting Auth errors throughout the day, Linux now seems OK?

Thank you, the connection seems to have drastically improved, thanks to all that helped fix the problem

@thread

Graze made some modifications last night. So far I've had a steady connection on Onyx which has lasted 8 hours, however Cantus still only lasted 25 mins so I'm not sure if the updates reached that server when I connected to it.

@Graze

I still get authentication error and disconnects on cantus (auth errors 3 today) and onyx (auth errors one today).

Win 7 64, widget 1.10

@kiwifruit

The issue is server-side, nothing on your end. The staff are spinning the mouse wheels as fast as they can.


@thread

OK, my connection to Onyx is at 15 hours and going strong, so I think the patch works (at least for that box).

I have one or two disconnects at PC and overall pretty slow performance on android for some reason. All with Cantus RAW.

The connection is still not good on fenrir. I am using Linux and it stays for 10-20 minutes.

@thread

My link to Cantus has been up just under 7 hours and counting. I'd say the patch worked for that node.


@Lunar

Fenrir has had other connectivity issues. It could be that the crew is fixing those first before applying the patch applied to the other nodes.

@thread

OK, cantus has been solid for about 16 hours now.

Shadow-raw
Currently 7+ hours uptime from pfSense router.

Hi,

up since 30 hours on cantus (Mon Jul 21 02:13:53 2014 -- Tue Jul 22 12:54:16 2014)

/Fermi

Bruno, which was previously good for me while the others were not is now dropping after about 20 minutes. I tried it 3 times and have now switched to onyx and will also use cantus. Thanks for the hard work and constant updates

@thread

OK, Cantus has passed the 24 hour mark. Gonna move on to Fenrir.

@thread

OK, quick update. My first connection to Fenrir lasted 20 mins. The second connection to Fenrir lasted 1hr26mins. In both cases I got this at the point of failure:

Using the Unofficial CryptoStorm Wiki I have managed to set up my connection to last 12+ hours. The connection is however much slower now than it was before all these troubles started.

@gbj

Which exit node did you get 12+ hours out of?

With the adress provided in iceland-fenrir-new.conf 79.134.235.133

I had this issue with the Iceland servers before but now I'm sitting pretty! Thanks for fixing it! and Hope everyone else's get's fixed soon too!

parityboy wrote:@thread

OK, Cantus has passed the 24 hour mark. Gonna move on to Fenrir.

I'm afraid the issues seem to be very hit and miss. I've been on cantus this morning and stayed connected for about 10 minutes the first time, 20 minutes the second. Back to bruno to see how it does. Has anyone confirmed these issues with openvpn? Maybe time to go back to that

EDIT: I can't stay connected to bruno for more than 6-8 minutes but my wife has been on bruno for the better part of 2 days on her own token. Is it possible these problems are related to individual tokens? Just a guess, but for certain I cannot stay connected.

I unfortunately have very mixed experiences with the nodes:

Germany has phases when it works flawlessly but then long periods of very sluggish performance and repeated disconnects.
Switching the node helps sometimes. Yesterday switching to France helped a lot but today it started failing the same way like Frankfurt did yesterday. In return Frankfurt seems to be more or less ok for now.

I also very clearly see that PCs connected via ethernet cable work way better than mobile devices connected via WIFI. The latter have phases where connecting to a node takes very long or fails all together. I was able to verify that with different WIFIs at home, at university and a public library. So it's neither my router nor my connection. I guess it has something to do with Android that reacts different than Ubuntu. (afaik we once had a situation where PC and Android reacted very different.)

I hope those issues can be resolved very soon. :\

kelltech,

I've been connected to cantus for +48 hours, without any issue.
The issue which has been tackled had as a symptom that the connection dropped every 20 minutes, resulting in an inactivity time out which leads to a reconnect.

The issue you encounter seems to be different from the initial issue.
Is it possible to post some logfiles?

/Fermi

@kelltech

Can you use your wife's token on your machine, as a test?

AUTH_FAILED on Maple raw and Bruno raw the last couple days. Here are portions of the logs from this morning. Both are the same other than the ip.

Wed Jul 23 09:24:20 2014 us=534625 Data Channel Encrypt: Cipher 'AES-256-CBC' initialized with 256 bit key
Wed Jul 23 09:24:20 2014 us=534666 Data Channel Encrypt: Using 512 bit message hash 'SHA512' for HMAC authentication
Wed Jul 23 09:24:20 2014 us=534703 Data Channel Decrypt: Cipher 'AES-256-CBC' initialized with 256 bit key
Wed Jul 23 09:24:20 2014 us=534729 Data Channel Decrypt: Using 512 bit message hash 'SHA512' for HMAC authentication
Wed Jul 23 09:24:20 2014 us=534782 UDPv4 WRITE [22] to [AF_INET]198.27.89.56:443: P_ACK_V1 kid=0 [ 42 ]
Wed Jul 23 09:24:20 2014 us=534874 Control Channel: TLSv1, cipher TLSv1/SSLv3 DHE-RSA-AES256-SHA, 2048 bit RSA
Wed Jul 23 09:24:22 2014 us=648805 SENT CONTROL [server]: 'PUSH_REQUEST' (status=1)
Wed Jul 23 09:24:22 2014 us=648921 UDPv4 WRITE [104] to [AF_INET]198.27.89.56:443: P_CONTROL_V1 kid=0 [ ] pid=11 DATA len=90
Wed Jul 23 09:24:22 2014 us=687002 UDPv4 READ [22] from [AF_INET]198.27.89.56:443: P_ACK_V1 kid=0 [ 11 ]
Wed Jul 23 09:24:22 2014 us=687126 UDPv4 READ [104] from [AF_INET]198.27.89.56:443: P_CONTROL_V1 kid=0 [ ] pid=43 DATA len=90
Wed Jul 23 09:24:22 2014 us=687245 AUTH: Received control message: AUTH_FAILED
Wed Jul 23 09:24:22 2014 us=687275 SIGTERM received, sending exit notification to peer
Process eventually exits.

Here's from Shadow raw, Shadow is working fine.
Wed Jul 23 09:36:48 2014 us=765524 Data Channel Encrypt: Cipher 'AES-256-CBC' initialized with 256 bit key
Wed Jul 23 09:36:48 2014 us=765565 Data Channel Encrypt: Using 512 bit message hash 'SHA512' for HMAC authentication
Wed Jul 23 09:36:48 2014 us=765591 Data Channel Decrypt: Cipher 'AES-256-CBC' initialized with 256 bit key
Wed Jul 23 09:36:48 2014 us=765617 Data Channel Decrypt: Using 512 bit message hash 'SHA512' for HMAC authentication
Wed Jul 23 09:36:48 2014 us=765671 UDPv4 WRITE [22] to [AF_INET]198.50.119.173:443: P_ACK_V1 kid=0 [ 42 ]
Wed Jul 23 09:36:48 2014 us=765765 Control Channel: TLSv1, cipher TLSv1/SSLv3 DHE-RSA-AES256-SHA, 2048 bit RSA
Wed Jul 23 09:36:48 2014 us=765840 [server] Peer Connection Initiated with [AF_INET]198.50.119.173:443
Wed Jul 23 09:36:50 2014 us=833682 SENT CONTROL [server]: 'PUSH_REQUEST' (status=1)
Wed Jul 23 09:36:50 2014 us=833790 UDPv4 WRITE [104] to [AF_INET]198.50.119.173:443: P_CONTROL_V1 kid=0 [ ] pid=11 DATA len=90
Wed Jul 23 09:36:50 2014 us=875084 UDPv4 READ [22] from [AF_INET]198.50.119.173:443: P_ACK_V1 kid=0 [ 11 ]
Wed Jul 23 09:36:50 2014 us=875201 UDPv4 READ [114] from [AF_INET]198.50.119.173:443: P_CONTROL_V1 kid=0 [ ] pid=43 DATA len=100
Wed Jul 23 09:36:50 2014 us=875301 UDPv4 WRITE [22] to [AF_INET]198.50.119.173:443: P_ACK_V1 kid=0 [ 43 ]
Wed Jul 23 09:36:50 2014 us=875400 UDPv4 READ [114] from [AF_INET]198.50.119.173:443: P_CONTROL_V1 kid=0 [ ] pid=44 DATA len=100
Wed Jul 23 09:36:50 2014 us=875483 UDPv4 WRITE [22] to [AF_INET]198.50.119.173:443: P_ACK_V1 kid=0 [ 44 ]
Wed Jul 23 09:36:50 2014 us=875558 UDPv4 READ [64] from [AF_INET]198.50.119.173:443: P_CONTROL_V1 kid=0 [ ] pid=45 DATA len=50
Wed Jul 23 09:36:50 2014 us=875728 PUSH: Received control message: 'PUSH_REPLY, ......... etc. etc

Log in is successful and connection stable.

Progress! Science! and all that.

I glad to see that some of you are back to seeing stable connections, and I have made sure our nerds behind the curtain have been made aware of all the information you all have provided. I'll dispense the updates here as I get them. Thanks so much for your continued patience and all the help everyone has provided in helping us get this issue tracked down.

@Fermi; @parityboy I've been able to stay connected all day today on emerald and cantus, going to leave it on fenrir overnight and then try bruno and cantus once more tomorrow.

Cantus is still not fine for me:

At daytimes the connection is speedy on PC and I only get very very few (if any) disconnects on Android.
Starting with the evening Cantus performance reduces very drastically and I get only slow speeds. I can rule out my normal internet connection as this one works perfectly fine without VPN.

Can you please check if Cantus for why it slows down every evening?

I use Cantus raw btw.

@thread

I think Cantus is playing up again, the connection seems to last about 1h20m before dropping. Will try to gather more data as I get it.

Hi,

I was unable to connect to Cantus this morning and a part of the afternoon. After contacting support things got solved. An hour ago connection dropped again, but I was able to reconnect.
Will follow up.

/Fermi

@thread

OK, the connection to Cantus has stayed stable since my last post, so I don't know what to make of things. Will continue to post new data as I get it.

Anyone still having some issues with cantus? I'm having someone take a look right now, and will report back with updates if it's warranted

Hi,

I haven't experienced any issues while using Cantus the last 24 hours.

/Fermi

@thread

Cantus just dropped for me, after staying up ~36 hours.

So, I've did some testing in Cantus with 2 different token's (Tk1 - 6 month and Tk2 - 1 month) both on PC, one with Windows 7 (PC A) and the other with Ubuntu 14.04 (PC B), from 11/08/2014 to 16/08/2014

- Both PC's run with the same wifi/network connection
- Both connected with 2 different token's
- Both running latest openvpn program

1) PC A with Tk1 - stayed connected until 14/08 but it reconnected after some seconds until 16/08
2) PC B with Tk2 - stayed connected until 12/08, then reconnected until 14/08, then reconnected until 16/08
3) Speedtest.net PC A - http://www.speedtest.net/result/3697628548.png no matter if it's day or night in UTC+01:00 country, this is the average
4) Speedtest.net PC B - http://www.speedtest.net/result/3697581501.png no matter if it's day or night in UTC+01:00 country, this is the average
5) Speedtest.net a long time ago it was like this http://www.speedtest.net/result/3438786795.png, a really big difference here

Hope this helps

Hi Tealc,

I've tested my connection with Cantus:
http://www.speedtest.net/my-result/3698938146
and
http://www.speedtest.net/my-result/3699028667
Results are of a different order than these you are experiencing.
I'm very happy with this performance. Could be that due to shutdown of bruno/shadow, some people are connecting to Cantus and thus impacting performance.

Regards,

/Fermi

One speedtest, later that day:
http://www.speedtest.net/my-result/3700220062

/Fermi