CrashPlan not connecting to CrashPlan Central

[Tookerder]

I'm wondering if anyone else has encountered this issue. Since updating to the latest 9.10 stable version, crashplan won't connect to CrashPlan central. Everything else about the plugin still works, and if I roll back the boot environment it connects again just fine. Also, all my incoming backups still function as well and so do DNS lookups. When i do a sockstat it isn't listening on 4243, so I think that is the issue possibly but not sure how to resolve. I tried putting the jail into dmz on my router temporarily and that did not help either.

As always, thanks in advance trusty freenas community!

Sent from my XT1254 using Tapatalk

 

[Nick2253]

Given the absolute cluster that is CrashPlan on *BSD, I would recommend moving your CrashPlan server to a Linux VM using the virtualbox plugin (this is what I do). If only CrashPlan officially supported FreeNAS (or at least FreeBSD), it would make a much happier world.

 

[Tookerder]

Thanks for the quick response. I have certainly considered that. I currently have a win10 vm running fine and an Ubuntu vm that is turned off. If you don't mind me asking, what linux version do you run crashplan on? Also, how much ram do you allocate? 2gb?

 

[Nick2253]

I use CentOS 6 because that is what I administer day-to-day at work, and I'm very familiar with it. I'm not backing up a ton of data, so I only allocate 1GB. I could probably get away with 512MB if I wanted to push it, but I'm not running in to a lot of memory bottlenecks, so it hasn't come up yet.

 

[Visseroth]

I am also having some problem...
I am trying to get Pro to work and I'm getting the error "Unable to connect to central.crashplan.com:443, check your network" when it should by trying to connect to central.crashplanpro.com:443

 

[millst]

It worked fine for me on 9.10 until very recently (last FreeNAS update?). I ended up putting the IP addresses in my jail's hosts file. It's an ugly hack, but Crashplan has been working since.

-tm

 

[Tookerder]

Yes, same for me. Was working until last update a few weeks ago. Which IPs did you put in the hosts file? CrashPlan? I don't have any trouble pinging CrashPlan from within the jail.

Sent from my XT1254 using Tapatalk

 

[millst]

DNS works fine in my jail, too. Look in the log file for the host resolution errors. They have the host name with port 443. Those need to be added.

-tm

 

[Visseroth]

I'm still a bit of a noob and I'm not sure where those files are located

 

[millst]

Start with the two addresses you already noted. I'd have to look through my system at home to give you more details, but I'm using a jail that I created (instead of the plugin) so things will be different. Running crashplan in an unsupported config is probably a bad idea for a self-described noob.

-tm

 

[Visseroth]

I'm half tempted to just install Crashplan in a linux vm in VirtualBox then share the folder via the Jail and then through VirtualBox. Yea, has to go through some layers but makes maintenance/updating way easier.
Problem is one of the other machines I'm working on (A TrueNAS Box) the VirtualBox seems to be broken. It gives me a error saying it can't connect to the local host when I try and log in. That's another issue but makes it hard to come up with a CrashPlan resolution when neither CrashPlan nor VirtualBox want to work.

 

[millst]

Yeah, I see a lot of people going that route lately. It seems Code42 are getting a little more hostile towards the unsupported OS crowd.

These are the two entries I added to my hosts file (based on errors in service.log.0)

aff-sea.crashplan.com
central.crashplan.com

-tm

 

[fracai]

Add me to the VM list. Actually the bhyve list. I created a bhyve vm, installed Debian, and then installed Crashplan and tightvnc. I'm sharing the data over NFS (mounted with autofs) so Crashplan still only notices new data every night when it verifies data, but being able to connect over VNC is far easier than trying to swap around configuration files. I haven't had any connectivity issues.
The one remaining task is getting bhyve to start the instance on boot.

 

[Tookerder]

I setup the VM as well, arch Linux. What is the advantage of bhyve over the VM jail plugin? I ended up sharing the datasets I'm backing up over cifs. Is NFS better do you think? I'm sort of new to all of this.

Sent from my XT1254 using Tapatalk

 

[fracai]

I don't know any concrete advantages. For now, bhvye requires manually creating and managing the instance via the command line. FreeNAS 10 should have GUI tools for this and should make things easier to keep it running. I also can't compare NFS and CIFS other than to say I have more experience with NFS and that's restricted to v3. I have not taken the time to figure out how to get NFSv4 working with permissions.
I would expect to bhyve to be a better experience in the end as it's more closely integrated with the host OS, but I can't speak directly to any performance advantage.

 

[dsfew32]

Not sure what it was but spent HOURS trying to find a solution (great idea about /etc/hosts that fixed it for me). But it's either 9.10 or 4.7 update or something but I'm getting errors like:

Code:

[06.30.16 01:04:35.713 WARN  DefaultGroup .code42.messaging.peer.PeerGroup] PG::DefaultGroup PeerException attempting to connect. RemotePeer-[guid=676143074786672641, SERVER]; PeerConnectionState-[state=CONNECTING, mode=PRIVATE, currentAddressIndex=0, layer=1/1(Peer::NIO), connecting=2016-06-30T01:04:35:712, connected=0, disconnected=2016-06-30T01:02:05:551, attempts=5, connectActivity=2016-06-30T01:04:35:712, keepAliveSent=0, minRetry=29778, reconnectable, location=eec-sea.crashplan.com:443, transportPbK=X509.checksum(8df6b6acb9425f1700804a8a423708ef), transportPbKRequestable=false]; Session-null e=com.code42.peer.exception.PeerException: IOExcepton opening remote session. guid=676143074786672641, remoteLocation=[[676143074786672641@eec-sea.crashplan.com:443(server)], transportPbK=X509.checksum(8df6b6acb9425f1700804a8a423708ef)], timeout=30000, java.io.IOException: Unexpected Exception in connect() for remoteAddress=eec-sea.crashplan.com:443, java.nio.channels.UnresolvedAddressException

These UnresolvedAddressException errors are found in various places with crashplan but even after reinstalling everything, messing with memory, resolv.conf, different versions, and everything I could find I couldn't fix it. Shoudl have checked here first.

 

[katit]

Same issue here. Really hate idea of layers and VirtualBox..

Millst,

Can you explain where those entries need to be added?

 

[katit]

I use CentOS 6 because that is what I administer day-to-day at work, and I'm very familiar with it. I'm not backing up a ton of data, so I only allocate 1GB. I could probably get away with 512MB if I wanted to push it, but I'm not running in to a lot of memory bottlenecks, so it hasn't come up yet.

Nick,

Do you run CentOS with GUI? I'm just asking because if not - it's same kind of setup as in jail but perhaps easier since CentOS is more widely used. But still need to use other computer for CrashPlan GUI..

 

[dsfew32]

Same issue here. Really hate idea of layers and VirtualBox..

Millst,

Can you explain where those entries need to be added?

From the jail terminal in the gui or ssh and run jls/jexec. Modify /etc/hosts following the format you see. For domain use whatever you had errors for in the error logs (I only had to add central.crashplan.com). For IP ping the domain and see what ip you got back. Crashplan is in some distribution/load balancing network it seems so there are many IPs for a given domain. You'll probably want the one you get back since it will be the closest to you.

 

[katit]

From the jail terminal in the gui or ssh and run jls/jexec. Modify /etc/hosts following the format you see. For domain use whatever you had errors for in the error logs (I only had to add central.crashplan.com). For IP ping the domain and see what ip you got back. Crashplan is in some distribution/load balancing network it seems so there are many IPs for a given domain. You'll probably want the one you get back since it will be the closest to you.

Sorry, I'm little bit lost here and need more advanced instructions :(
1. When I SSH to crashplan instance - running "jls/jexec" doesn't do anything (command not found)
2. I found etc/hosts but not sure where error log is for crashplan? I only found in /var/log/crashplan and there is 2 files with not much info..