CrashPlan Plugin

[afmiller]

I'm getting hung up on the putty section of the tunnel now.

I verified my settings
Before You Begin

• Be sure CrashPlan is running on your remote machine.
• Verify on the headless machine (with netstat) that it is listening on port 4242 on all addresses and on port 4243 on the local address (this is the UI service port).

netstat -na | grep LISTEN | grep 42
Output should look like this:
username$ netstat -na | grep LISTEN | grep 42
tcp4 0 0 *.4242 *.* LISTEN
tcp4 0 0 127.0.0.1.4243 *.*
We want to use SSH to tunnel a local Windows port (4200) to the remote host's service port (4243).
This is the output after I follow the instructions for the putty part. I am using putty as user1 , which is tied to the crashplan jail. I am able to putty into the Jail via 192.168.1.2 , but after I do the telnet localhost:4200 I get the below output

$ telnet 192.168.1.2 4200
Trying 192.168.1.2...
telnet: connect to address 192.168.1.2: Connection refused
telnet: Unable to connect to remote host

 

[afmiller]

got it working

 

[David E]

Ah, that's the part I was missing, though the config change went effective instantly, restarted the service and I'm able to SSH into it now. Is it normal for the /mnt drive to show as 1 file and 0 mb in the crashplan app?

Yes, you need to click the "Add Storage" button for the crashplan_1 jail, which will mount your file systems inside the jail so you can back them up.

 

[didolgi]

Crashplan is releasing new code this week. I found that once my other PCs were updated to the new code they stoped talking to the CrashPlan on the FreeNAS box. I don't know if the CrashPlan inside the jail would have updated itself to the new code or not. I did a manual update to the new code, and now my PCs are all talking to the updated version just fine. I did a manual install to a 'standard jail' following the steps here - http://forums.freenas.org/index.php...an-in-freenas-9-1-jail-tested-on-32bit.14099/

Following those steps on a FreeNAS 9.2.1-RELEASE(x64) worked just fine.

 

[David E]

Anyone else seeing these errors in their nightly emails now:

linux: pid 50989 (java): syscall epoll_create not implemented

I'm not sure if the JVM is smart enough to fall back from epoll to poll or select if the call fails, if anyone else is seeing these errors, are your backups still working properly?

 

[Seany]

I'm also getting this error (syscall epoll_create not implemented) after having no problems with the plugin for a couple weeks.

I just recently resilvered, not sure if that has anything to do with it.

 

[Mark Hodgkinson]

I did add the mount points. I have 4 of them filled up from mounting to my pool for my family photos, docs, videos, and music.

While I followed Sully to the letter (e.g., no VIMAGE, yes, NAT), I turn it on, SSH in from a client and manually start the CrashPlanService OK, but I am getting:

Java HotSpot(TM) Client VM warning: Can't detect initial thread stack location - find_vma failed
In the engine_output.log and
/etc/host.conf: line 2: bad command `hosts'
/etc/host.conf: line 3: bad command `dns'
from the engine_error.log
Not sure why, as I mod'd root's pwd inside the jail and restarted that service so I could SSH in from client just fine.

I'm getting the same error since Crashplan updated itself but I don't understand how you fixed it with:

I finaly found what was the problem.

If anyone wants to know the reason, here it is:
it needs /proc !
so, you can just do:
mount --bind /proc /chrootedpath/proc

What should the chrootedpath be set to?

 

[Mark Hodgkinson]

Cancel that, I've managed to get it working after a reboot, no idea how though!

 

[Seany]

Looks like my problem is related to crashplan attempting to upgrade from 3.5.3 to 3.6.3.
Even when I uninstall and reinstall the plugin, the engine attempts to upgrade when I connect to it from the Crashplan desktop app, and stops working.

 

[shawndewet]

Ok so I get my CrashPlan plugin working fine. I can connect my crashplan client to the crashplan service running on my FreeNAS. I get my first backup configured with crashplan central as the destination. Backup status shows "synchronising".
As per Seany's post just here before mine, I also noticed a message about a new version being downloaded and installed, and that the UI will be temporarily unavailable, and it exits.
Later I want to check the status...and the client won't connect - "Unable to connect tot he backup engine, retry?"
So I look at my FreeNAS plugins and indeed the CrashPlan plugin is NOT running. I click the button, and it takes a few seconds and switches from Off to On. To confirm, I refresh the tab, and, to my disappointment, it is back to the Off position. NOT running.
What can I check to see why the service is not starting (even though the FreeNAS UI does flip the On/Off switch to On)?

 

[mstinaff]

I have a solution here that works. I found it after conencting a 3.5.3 gui to my plug in and it auto updated. I then tried a fresh plugin and connected with a 3.6.3 gui. Gave some message about setting a backup password, but worked and did not update the engine for a while.

Last week the engine did update itself and the above fix got the engine running again. One gotcha though, the error
Java HotSpot(TM) Client VM warning: Can't detect initial thread stack location - find_vma failed
that marcevan noted is still present. But so far it only seems to mean that running crashplan status when sshed in will say that the engine is not running even if it is. I still get this error, but the engine runs fine and I can connect with a 3.6.3 gui through a tunnel and monitor the progress.

 

[sully]

I have a solution here that works. I found it after conencting a 3.5.3 gui to my plug in and it auto updated. I then tried a fresh plugin and connected with a 3.6.3 gui. Gave some message about setting a backup password, but worked and did not update the engine for a while.

Last week the engine did update itself and the above fix got the engine running again. One gotcha though, the error
Java HotSpot(TM) Client VM warning: Can't detect initial thread stack location - find_vma failed
that marcevan noted is still present. But so far it only seems to mean that running crashplan status when sshed in will say that the engine is not running even if it is. I still get this error, but the engine runs fine and I can connect with a 3.6.3 gui through a tunnel and monitor the progress.

I tried the above solution but wasn't able to get the plugin running. I got the following errors.

When I check the status of the engine...

root@crashplan_1:/ # service crashplan status

/usr/local/etc/rc.d/crashplan: WARNING: no shebang line in /usr/bin/cpuset

btw, I laughed when I saw the "shebang"... wth is that!?

Error in the activity..

Mar 5 12:49:26 armory kernel: linux: pid 83441 (java): syscall epoll_create not implemented
Mar 5 12:49:37 armory kernel: linux: pid 83493 (java): syscall epoll_create not implemented
Mar 5 12:50:18 armory kernel: linux: pid 83637 (java): syscall epoll_create not implemented
Mar 5 12:50:38 armory kernel: linux: pid 83689 (java): syscall epoll_create not implemented

I know enough to be dangerous, not enough to resolve this issue though.

 

[mstinaff]

Try running this in your jail after editing the file

Code:

root@crashplan_1:/ # /usr/bin/cpuset -l 0 /usr/local/share/crashplan/bin/CrashPlanEngine restart  

Thats cpuset with dash lowercase L and then number zero

Then do

Code:

root@crashplan_1:/ # ls /proc 

you should see a fistful of entries.


Then run

Code:

root@crashplan_1:/ # sockstat -4 

you should see a bunch of entries for java

p.s.
shebang is the #! at the start of scripts

 

[sully]

Try running this in your jail after editing the file

Code:

root@crashplan_1:/ # /usr/bin/cpuset -l 0 /usr/local/share/crashplan/bin/CrashPlanEngine restart  

Thats cpuset with dash lowercase L and then number zero

Then do

Code:

root@crashplan_1:/ # ls /proc 

you should see a fistful of entries.


Then run

Code:

root@crashplan_1:/ # sockstat -4 

you should see a bunch of entries for java

p.s.
shebang is the #! at the start of scripts

Result

Code:

root@crashplan_1:/ # /usr/bin/cpuset -l 0 /usr/local/share/crashplan/bin/CrashPlanEngine restart
Restarting CrashPlan Engine ... Stopping CrashPlan Engine ... /usr/pbi/crashplan-amd64/share/crashplan/bin/CrashPlanEngine: line 145: kill: (91713) - No such process
OK
Starting CrashPlan Engine ... Using standard startup
OK
root@crashplan_1:/ # ls /proc
88212 88265 88469 92446 92549 curproc
root@crashplan_1:/ # sockstat -4
USER COMMAND PID FD PROTO LOCAL ADDRESS FOREIGN ADDRESS
root python2.7 88469 3 tcp4 192.168.2.201:12348 *:*
root syslogd 88212 6 udp4 192.168.2.201:514 *:*
 
root@crashplan_1:/ #

I appreciate your help.

 

[mstinaff]

Looks like it might be spinning up. Could you post the contents of your /var/log/crashplan/engine_output.log

If it is running the first line should be
Java HotSpot(TM) Client VM warning: Can't detect initial thread stack location - find_vma failed

and the last line should be
jtux loaded

 

[sully]

Here's the engine_output.log Loads, then quits, Java related?

Code:

Java HotSpot(TM) Client VM warning: Can't detect initial thread stack location - find_vm
[03.05.14 13:26:59.443 INFO    main                root
[03.05.14 13:26:59.443 INFO    main                root
[03.05.14 13:26:59.443 INFO    main                root
[03.05.14 13:26:59.443 INFO    main                root
[03.05.14 13:26:59.444 INFO    main                root
[03.05.14 13:26:59.444 INFO    main                root
[03.05.14 13:26:59.445 INFO    main                root
[03.05.14 13:26:59.445 INFO    main                root
[03.05.14 13:26:59.509 INFO    main                root
[03.05.14 13:26:59.526 INFO    main                root
[03.05.14 13:26:59.553 INFO    main                root
[03.05.14 13:26:59.553 INFO    main                root
[03.05.14 13:26:59.553 INFO    main                root
[03.05.14 13:26:59.553 INFO    main                root
[03.05.14 13:26:59.553 INFO    main                root
[03.05.14 13:26:59.561 INFO    main                root
[03.05.14 13:26:59.609 INFO    main                root
[03.05.14 13:26:59.695 INFO    main                root
[03.05.14 13:26:59.695 INFO    main                root
[03.05.14 13:26:59.818 INFO    main                root
[03.05.14 13:26:59.819 INFO    main                root
[03.05.14 13:26:59.820 INFO    main                root
[03.05.14 13:26:59.821 INFO    main                root
jtux Loaded.
Exiting!!! java.lang.RuntimeException: Unexpected IO Exception constructing selector eng
[03.05.14 13:27:00.447 INFO    main                root
[03.05.14 13:27:00.448 INFO    Thread-0            root
[03.05.14 13:27:00.460 INFO    Thread-0            root
[03.05.14 13:27:00.461 INFO    Thread-0            root
[03.05.14 13:27:00.461 INFO    Thread-0            root
 
 

 

[mstinaff]

can you post the contents of you /usr/pbi/crashplan-amd64/share/crashplan/bin/run.conf

 

[sully]

can you post the contents of you /usr/pbi/crashplan-amd64/share/crashplan/bin/run.conf

Ok, that's interesting. I know for sure I made the change to the run.conf. I just looked at it again and the changed didn't take, why I have no clue. Maybe it was reverted somehow? Either way, I again added the lines from http://forums.freenas.org/index.php?threads/crashplan-3-6-3.18416/#post-101205 and restarted the service. I'm now able to connect to the engine, no java errors in the console.

Going to run the backup, restart the jail/plugin and see the added lines stick.

Thanks again for your help mstinaff.

 

[David E]

There is some strange sticking problem, I recommend shutting down the jail and changing the file via FreeNAS, then starting up, that worked for me.

 

[michiganblue82]

I just installed Crashplan 3.6.3_1 on FreeNas 9.2.1 and am getting the error "Crashplan did not validate, configure it first". I have no idea where to start. Any help would be much appreciated. Thank you!