Crashplan plugin 3.6.3 service won't start after system reboot

[Halfe]

Ive been trying to solve this problem for a couple of days now and found out that the run.conf file returns to orginal. deleting the edited lines after reboot.
SRV_JAVA_OPTS="-Djava.nio.channels.spi.SelectorProvider=sun.nio.ch.PollSelectorProvider -Dfile............"
GUI_JAVA_OPTS="-Djava.nio.channels.spi.SelectorProvider=sun.nio.ch.PollSelectorProvider -Dfile............"

But when i got the Engine working, i was unable to connect to network. Been checking my Firewall but there is no indication that Jail IP is trying to connect to central.crashplan.com.

I pinged Central.crashplan.com from within jail and got results back

Code:

crashplan@crashplan_1:~ % ping central.crashplan.com
PING central.crashplan.com (216.17.8.51): 56 data bytes
64 bytes from 216.17.8.51: icmp_seq=0 ttl=51 time=145.036 ms
64 bytes from 216.17.8.51: icmp_seq=1 ttl=51 time=144.384 ms
64 bytes from 216.17.8.51: icmp_seq=2 ttl=51 time=143.977 ms
64 bytes from 216.17.8.51: icmp_seq=3 ttl=51 time=144.523 ms

testing CP on my PC works great. got it working by opening it up in the firewall.

 

[bcknapp]

Thank you to everyone who has responded.

It looks like perhaps there was an update as my run.conf file and my.service.xml file were reverted to default values. I added back in the lines that were referenced in this post and upped the memory allocation to 4096m which is what I was previously doing when I was running the service in Windows (this solved a problem where the service stopped backing up). But when I run the sockstat command I do not see any entries for java. This is after shutting rebooting the jail, and serveral restarts of the entire freenas box. On the plugin list, I cannot get the plugin to say "on." I wonder what is going on because it sounds like this fixed the problem for some of you.

This is still some of the output I am getting:

Code:

root@crashplan_1:/ # service crashplan status                                  
/usr/local/etc/rc.d/crashplan: WARNING: no shebang line in /usr/bin/cpuset     
crashplan is not running.     

Code:

root@crashplan_1:/ # sockstat                                                  
USER     COMMAND    PID   FD PROTO  LOCAL ADDRESS         FOREIGN ADDRESS      
root     python2.7  34805 3  tcp4   192.168.1.2:12346     *:*                  
root     cron       34581 4  dgram  -> /var/run/logpriv                        
root     syslogd    34514 4  dgram  /var/run/log                               
root     syslogd    34514 5  dgram  /var/run/logpriv                           
root     syslogd    34514 6  udp6   *:514                 *:*                  
root     syslogd    34514 7  udp4   *:514                 *:*                 

Code:

root@crashplan_1:/ # ls /proc                                                  
34514   34581   34805   46591   46699   curproc                              

This is the contents of my run.conf file:

Code:

SRV_JAVA_OPTS="-Djava.nio.channels.spi.SelectorProvider=sun.nio.ch.PollSelectorP
rovider -Dfile.encoding=UTF-8 -Dapp=CrashPlanService -DappBaseName=CrashPlan -Xm
s20m -Xmx4096m -Djava.net.preferIPv4Stack=true -Dsun.net.inetaddr.ttl=300 -Dnetw
orkaddress.cache.ttl=300 -Dsun.net.inetaddr.negative.ttl=0 -Dnetworkaddress.cach
e.negative.ttl=0 -Dc42.native.md5.enabled=false"                               
GUI_JAVA_OPTS="-Djava.nio.channels.spi.SelectorProvider=sun.nio.ch.PollSelectorP
rovider -Dfile.encoding=UTF-8 -Dapp=CrashPlanDesktop -DappBaseName=CrashPlan -Xm
s20m -Xmx512m -Djava.net.preferIPv4Stack=true -Dsun.net.inetaddr.ttl=300 -Dnetwo
rkaddress.cache.ttl=300 -Dsun.net.inetaddr.negative.ttl=0 -Dnetworkaddress.cache
.negative.ttl=0 -Dc42.native.md5.enabled=false"    

Still looking for a solution because, as someone else stated, I am really enjoying the functionality of my FreeNAS box, but not having CrashPlan is kind of a deal breaker. Thanks again for all the replies!

 

[danb35]

In one of the related threads, I've seen the suggestion that it won't accept a memory limit over about 3 GB. You might want to try that.

 

[bcknapp]

danb35,

Thanks for letting me know. I've set it back to 2048. I am still getting the same results, however.

 

[bcknapp]

Ok, I have just looked back at my run.conf file for about the 10th time in the past hour, and I see that it has completely reverted back again, so I added in the string at the beginning and set the memory to 2048, restarted the plugin, and now the CrashPlan client connects! I don't know why it took so many times or what combination of turning off/on the plugin, turning off/on the jail and restarting the freenas machine did it, but it looks like it is running, for now!

Thanks to everyone who contributed to this discussion. At least I know that if/when this happens again I can probably getting back running again without having to delete the plugin/jail and start from scratch.

 

[Jailer]

You have to edit the file using ee, the changes won't stay if you try to edit it directly via ftp.

 

[bcknapp]

You have to edit the file using ee, the changes won't stay if you try to edit it directly via ftp.

That is an interesting point to note. Thank you for mentioning that. I have been editing using ee and the jail console from within the FreeNAS GUI. Not sure why the changes were not taking effect or sticking.

 

[fracai]

I have an idea that could help for the future. The rc startup script should execute a sed script that ensures the additional arguments are present in run.conf before starting crashplan. This should be really simple; something like:

Code:

's/ *-Djava.nio.channels.spi.SelectorProvider=sun.nio.ch.PollSelectorProvider *//;s/OPTS="/OPTS="-Djava.nio.channels.spi.SelectorProvider=sun.nio.ch.PollSelectorProvider /'

It basically removes the SelectorProvider section if it's there (including any leading or trailing space) and then adds the section back in at the start of the argument list. I haven't tested this so use with caution.

 

[Biggeek]

That is an interesting point to note. Thank you for mentioning that. I have been editing using ee and the jail console from within the FreeNAS GUI. Not sure why the changes were not taking effect or sticking.

I'm having the same issue.
I edit the files in the jail shell and they don't stick after reboot.
What could be changing them back?

 

[Biggeek]

Update-
I edited the file again, rebooted, started service, connected crashplan GUI.
Service is up and running, but from the jail shell, shows service is not running.
ODD.

 

[alexg]

I'm seeing same issue. Crashplan is up and running, but FreeNAS plugin shows service is down.

 

[Pasquale61]

I'm seeing same issue. Crashplan is up and running, but FreeNAS plugin shows service is down.

This is what I was getting after fixing run.conf and restarting several times. CrashPlan was running and working fine, but the FreeNAS GUI incorrectly showed that it was down. For some reason, after the latest 9.3 update and a complete reboot of everything, the plugin now correctly shows that the service is running.

 

[bcknapp]

I have an idea that could help for the future. The rc startup script should execute a sed script that ensures the additional arguments are present in run.conf before starting crashplan. This should be really simple; something like:

Code:

's/ *-Djava.nio.channels.spi.SelectorProvider=sun.nio.ch.PollSelectorProvider *//;s/OPTS="/OPTS="-Djava.nio.channels.spi.SelectorProvider=sun.nio.ch.PollSelectorProvider /'

It basically removes the SelectorProvider section if it's there (including any leading or trailing space) and then adds the section back in at the start of the argument list. I haven't tested this so use with caution.

Definitely an interesting idea. Might be a good solution in the meantime while this issue hopefully is getting some attention. I haven't tried something like that yet and I'm still new to the platform so I will probably not be the first to try it. If anyone gets it working, I'd certainly be interested to hear about it though.

I'm seeing same issue. Crashplan is up and running, but FreeNAS plugin shows service is down.

In the same situation here. Crashplan is running, but FreeNAS says the plugin is off.

 

[Jailer]

Well mine was running for a couple days but it's down now. I did check the run.conf file and it hasn't changed but it just won't run now.

 

[alexg]

After I rebooted my FreeNAS, plugin came up and has been staying in "on" position for two days now.

 

[bcknapp]

After I rebooted my FreeNAS, plugin came up and has been staying in "on" position for two days now.

I rebooted today, and now the plugin shows on for me as well, and the client connects with no trouble.

 

[Andrew076]

I seem to be having the same issue (or similar at least). My Crashplan (version 3.6.3_1) Plugin shows the Service status as "Off". Additional the Crashplan software says it is "Unable to connect to the backup enginge, retry?" (version 4.2.0).

I have tried going to the shell and typing service crashplan onestart to see if that fixed it. Then after that I ran service -e to check the status of what was enabled and it listed /user/local/etc/rc.d/crashplan which to me means that the service IS enabled. However something has changed as I can't get the software to connect now.

 

[fracai]

They updated to 4.2 this week. You probably have to reset the PollSelectorProvider value again.

 

[Andrew076]

How do I "rest the PollSelectorProvider value"

 

[Victor866]

The same happens to me, I only need to replace again the file ui.properties with plugin IP