Vynce Posted October 6, 2016 Share Posted October 6, 2016 There are a couple of errors in the sed commands in 01_config.sh which are causing Crashplan to be configured with bad ports and makes it unreachable. ${TCP_PORT_4242} on line 46 needs to be replaced with ${SERVICE_PORT} and ${TCP_PORT_4243} on line 47 needs to be replaced with ${BACKUP_PORT} Quote Link to comment
Helmonder Posted October 6, 2016 Share Posted October 6, 2016 I also notice that one of my two crashplans reports no internal ip address.. Weird thing is that they are both running the same unraid and same crashplan version.. Ofcourse both unraid webgui's are perfectly reachable on their respective internal addresses.. Hope this gets fixed soon.. I am missing my backups.. Quote Link to comment
Helmonder Posted October 6, 2016 Share Posted October 6, 2016 On the unraid server that its crashplan is now reporting a 0.0.0.0 address as local address (see above) I also have a win10 vm running with crashplan installed, it also shows a 0.0.0.0 address for the local IP... This crashplan version is ofcourse not a docker version but a windows version.. I therefor think that it is -not- the docker itself causing the issue.. That would mean it is either crashplan itself (but we should expect a lot to be google-able on that if crashplan itself would be borked, so i do not think that is it). Or there is some relation with the docker engine running on unraid. I have been having backup troubles for approx. 2 days which is about the same time unraid 6.2.1 is out.. Quote Link to comment
Hoopster Posted October 6, 2016 Share Posted October 6, 2016 I have been having backup troubles for approx. 2 days which is about the same time unraid 6.2.1 is out.. Hmmm, interesting theory. I updated to both the latest CrashPlan docker and unRAID 6.2.1 two days ago as well. I don't recall which update I did first. If others having troubles with connectivity to the CrashPlan engine are still on unRAID 6.2.0 then the latest CrashPlan docker is the culprit. If the only ones having trouble are all running unRAID 6.2.1 perhaps something changed there to affect connectivity with everything but the local GUI. I have looked in all the usual places (.ui.info, my.services, xml, .identity) well-documented in this forum for the reason why the Windows client can't connect and none of those seem to be the issue. Quote Link to comment
landS Posted October 6, 2016 Share Posted October 6, 2016 I am on CP 4.8 and UR 6.2.0 I pass through Mnt/User and backup a handful of shares Both accessing the GUI and backing up to the cloud is working fine. Quote Link to comment
tigerdoc Posted October 6, 2016 Share Posted October 6, 2016 Has anyone had any issues since last night? Specifically, CrashPlan is stopping and restarting every minute or so. I am able to connect via web GUI and see the CP history showing entries starting CP 4.8.0 with the correct GUID and scanning for files. Those entries repeat about every minute, since 12:00 am today. I'm running CP 4.8.0 and Unraid 6.2.1. Thanks. Quote Link to comment
Hoopster Posted October 6, 2016 Share Posted October 6, 2016 I am on CP 4.8 and UR 6.2.0 I pass through Mnt/User and backup a handful of shares Both accessing the GUI and backing up to the cloud is working fine. I am running latest CP 4.8 docker and unRAID 6.2.1 I pass through /mnt/user and backup several shares Both accessing the GUI and backing up to the cloud is working well (backup is running now) My issue is that since the update to either the latest CP docker or unRAID 6.2.1 (following Helmonder's theory), the CP Windows client will not connect to the CP backup engine. In Helmonder's case he also has connectivity issues between CP servers. Quote Link to comment
landS Posted October 6, 2016 Share Posted October 6, 2016 @Hoopster... Well... that narrows it down a bit more I suppose... beyond my setup to help though... @tigerdoc... I have about a dozen destination disconnects early last night... which is not the norm... however my backup is working fine. Perhaps try stopping the CP docker and restarting? Quote Link to comment
DDock Posted October 6, 2016 Share Posted October 6, 2016 I just started using the CrashPlan docker. I am trying to use it to backup my cache drive to my array. I saw you can make scripts in unRAID's go file to do it, but I'd rather have a bit more control. Right now when I try to open the docker (vnc) or try to VNC into it using TightVNC I am just getting a black screen. Is there something I have configured wrong? I have not updated to 6.2.1, still on 6.2. Quote Link to comment
tigerdoc Posted October 6, 2016 Share Posted October 6, 2016 @tigerdoc... I have about a dozen destination disconnects early last night... which is not the norm... however my backup is working fine. Perhaps try stopping the CP docker and restarting? Thanks for the suggestion. Tried stopping and restarting CP --> no difference. Forced update and restarted CP --> no difference. CP appears to crash and restart about every 60 seconds. If I'm connected to webUI at the time, VNC says it lost the connection and then a few seconds later shows CP restarting. Quote Link to comment
landS Posted October 6, 2016 Share Posted October 6, 2016 I just started using the CrashPlan docker. I am trying to use it to backup my cache drive to my array. I saw you can make scripts in unRAID's go file to do it, but I'd rather have a bit more control. Right now when I try to open the docker (vnc) or try to VNC into it using TightVNC I am just getting a black screen. Is there something I have configured wrong? I have not updated to 6.2.1, still on 6.2. Go to Docker, Click on the Play button, select WebUI Quote Link to comment
DDock Posted October 6, 2016 Share Posted October 6, 2016 That is what I meant by "open the docker", it opens it's own VNC viewing application. Still only shows a black screen. Quote Link to comment
Vynce Posted October 6, 2016 Share Posted October 6, 2016 What I was seeing is that the main Crashplan port was getting set to 0 and the service port set to 1 by 01_config.sh whenever the container was started. This would not occur if you set the TCP_PORT_4242 and TCP_PORT_4243 variables in the docker config. I did not have those variables set. grep -nHE '(location|servicePort)' /mnt/cache/appdata/crashplan/conf/my.service.xml /mnt/cache/appdata/crashplan/conf/my.service.xml:9: <location>0.0.0.0:0</location> /mnt/cache/appdata/crashplan/conf/my.service.xml:18: <servicePort>1</servicePort> I launched a bash shell in the container: " docker exec -it CrashPlan bash " and made the edits to 01_config.sh as described in my previous post. This allows Crashplan to start with the correct default ports and I'm able to connect to it again now. Quote Link to comment
landS Posted October 6, 2016 Share Posted October 6, 2016 That is what I meant by "open the docker", it opens it's own VNC viewing application. Still only shows a black screen. Have you tried a different Web-browser? I've had this happen a few times after browser updates and falling back to firefox when chrome is fussy has always worked. Quote Link to comment
benstrousers Posted October 6, 2016 Share Posted October 6, 2016 Thanks for the suggestion. Tried stopping and restarting CP --> no difference. Forced update and restarted CP --> no difference. CP appears to crash and restart about every 60 seconds. If I'm connected to webUI at the time, VNC says it lost the connection and then a few seconds later shows CP restarting. I'm having the same problem, these are showing as errors in the log. Running Unraid 6.2.1 and the latest commit of the Crashplan container. Removed container and image, reinstalled with same result. ERROR: openbox-xdg-autostart requires PyXDG to be installed ./run: line 20: 23237 Killed $JAVACOMMON $SRV_JAVA_OPTS -classpath "$TARGETDIR/lib/com.backup42.desktop.jar:$TARGETDIR/lang" com.backup42.service.CPService > /config/log/engine_output.log 2> /config/log/engine_error.log /opt/startapp.sh: line 22: 23374 Killed ${JAVACOMMON} ${GUI_JAVA_OPTS} -classpath "./lib/com.backup42.desktop.jar:./lang:./skin" com.backup42.desktop.CPDesktop > /config/log/desktop_output.log 2> /config/log/desktop_error.log Quote Link to comment
DDock Posted October 6, 2016 Share Posted October 6, 2016 I just recently got home from work. I was VPN'd into my home network working on all this. I tried to use the CrashPlan docker web GUI here and it still did not work. But if you remember from my previous post, a connection via TightVNC resulted in the same. I have uninstalled and reinstalled it probably a half dozen times, messing with different attributes each time with no luck. At this point I have no idea what is going on. I tried to look into the logs and nothing has been written into the xxx_error.log(s). Does having data already in the backup destination cause it to fail? I have my mounts like this: /unraid /mnt/cache (also tried /mnt/user) /unassigned /mnt/disks/ /config /mnt/cache/appdata/CrashPlan backup destination is /mnt/user/cache-backup Quote Link to comment
ljm42 Posted October 7, 2016 Share Posted October 7, 2016 Has anyone had any issues since last night? Specifically, CrashPlan is stopping and restarting every minute or so. I am able to connect via web GUI and see the CP history showing entries starting CP 4.8.0 with the correct GUID and scanning for files. Those entries repeat about every minute, since 12:00 am today. I'm running CP 4.8.0 and Unraid 6.2.1. Thanks. I figured out why CrashPlan keeps restarting - it needs more memory. I think the 4.8 upgrade wiped out our previous adjustments to run.conf This article explains the problem. Unfortunately the article no longer contains the details on how to solve it, it just says to contact support. So the fix is: shutdown the CP docker SSH to your server type: cd /mnt/cache/appdata/CrashPlan/bin (or wherever your CP appdata files are) type: nano run.conf on the line that starts with SRV_JAVA_OPTS, change "-Xmx1024m" to "-Xmx2048m" press CTRL-O, CTRL-X type: exit start the CP docker This doubles the amount of RAM available to CP from 1 GB to 2 GB. The article above explains what CP recommends, you may need a different number. Note: one thing I do is have multiple backup sets, I *think* (not positive) that this reduces the memory needs of CP. Unfortunately, it looks like the run.conf file can revert to defaults when CP upgrades (maybe other times too) gfjardim - would you consider adding a new env variable for "ram" which automatically updates the run.conf file when the container starts? Quote Link to comment
DDock Posted October 7, 2016 Share Posted October 7, 2016 Another update -- I have tried increasing the memory (since there are a few issues with this build it seems) but that did not help. Here is an oddity... While using TightVNC and the built in Docker UI still have a black screen, but I do have a mouse! If I move the mouse on one screen, the other moved, but very slowly and jerky, but it moves. Add that to the list. Quote Link to comment
Hoopster Posted October 7, 2016 Share Posted October 7, 2016 What I was seeing is that the main Crashplan port was getting set to 0 and the service port set to 1 by 01_config.sh whenever the container was started. This would not occur if you set the TCP_PORT_4242 and TCP_PORT_4243 variables in the docker config. I did not have those variables set. grep -nHE '(location|servicePort)' /mnt/cache/appdata/crashplan/conf/my.service.xml /mnt/cache/appdata/crashplan/conf/my.service.xml:9: <location>0.0.0.0:0</location> /mnt/cache/appdata/crashplan/conf/my.service.xml:18: <servicePort>1</servicePort> I launched a bash shell in the container: " docker exec -it CrashPlan bash " and made the edits to 01_config.sh as described in my previous post. This allows Crashplan to start with the correct default ports and I'm able to connect to it again now. Thanks Vynce, that did it. I edited 01_config.sh as you suggested and my Windows client again connected to the CrashPlan server. Of course, still had to do the copy .ui.info trick. The next CrashPlan Docker update will probably clobber the edits, but, at least I can connect through Windows client again. That is something I never would have found on my own. Perhaps setting those variables in the docker config is expected now by gfjardim. I noticed the newly generated .ui.info on the server had a "1" in place of the usual 4243 port number before the edits, so, that, confirms what you found as well. Quote Link to comment
Helmonder Posted October 7, 2016 Share Posted October 7, 2016 For the benefit of the rest: 1) log into your unraid server 2) log into the crashplan docker: docker exec -it CrashPlan bash 3) go to /etc/my_init.d 4) vi 01_config.sh ${TCP_PORT_4242} on line 46 needs to be replaced with ${SERVICE_PORT} ${TCP_PORT_4243} on line 47 needs to be replaced with ${BACKUP_PORT} And indeed, then it works ! Just restart the docker Quote Link to comment
tigerdoc Posted October 7, 2016 Share Posted October 7, 2016 @ljm42 -- Thanks for the memory tip. I updated to 3 GB and it has been running now for about a half hour without crashing. That may have solved it, at least for me. @Helmonder -- If you're using the new Appdata configuration, you'll find it at /mnt/user/Appdata/crashplan/bin/run.conf Quote Link to comment
izarkhin Posted October 7, 2016 Share Posted October 7, 2016 I have been having backup troubles for approx. 2 days which is about the same time unraid 6.2.1 is out.. Hmmm, interesting theory. I updated to both the latest CrashPlan docker and unRAID 6.2.1 two days ago as well. I don't recall which update I did first. If others having troubles with connectivity to the CrashPlan engine are still on unRAID 6.2.0 then the latest CrashPlan docker is the culprit. If the only ones having trouble are all running unRAID 6.2.1 perhaps something changed there to affect connectivity with everything but the local GUI. I have looked in all the usual places (.ui.info, my.services, xml, .identity) well-documented in this forum for the reason why the Windows client can't connect and none of those seem to be the issue. I'm having connectivity issues for 3 days now. Haven't been able to figure out how to fix it. Anybody have ideas how to troubleshoot? Quote Link to comment
Hoopster Posted October 7, 2016 Share Posted October 7, 2016 I'm having connectivity issues for 3 days now. Haven't been able to figure out how to fix it. Anybody have ideas how to troubleshoot? See Helmonder's post a couple before yours. We both did what Vynce suggested (edit 01_config.sh) several posts back and it resolved our connectivity issues with everything but the CrashPlan WebGUI (which works fine without the edits). Quote Link to comment
gfjardim Posted October 7, 2016 Author Share Posted October 7, 2016 For the benefit of the rest: 1) log into your unraid server 2) log into the crashplan docker: docker exec -it CrashPlan bash 3) go to /etc/my_init.d 4) vi 01_config.sh ${TCP_PORT_4242} on line 46 needs to be replaced with ${SERVICE_PORT} ${TCP_PORT_4243} on line 47 needs to be replaced with ${BACKUP_PORT} And indeed, then it works ! Just restart the docker Thanks for finding the bug. Container updated. Quote Link to comment
landS Posted October 7, 2016 Share Posted October 7, 2016 Mmmm.... I wonder if these 2 methods of increasing the Crashplan Ram/Memory differ in any way... Is it OK to increase the RAM from the default 1024 MB to 2048? As per: Open the CrashPlan app Double-click the CrashPlan logo in the upper-right corner Enter the following command, using a value appropriate for the size of your backup selection (for example, 1536 for a 1.5 TB selection): java mx 1536, restart The CrashPlan app closes http://support.code42.com/CrashPlan/4/Troubleshooting/Adjusting_CrashPlan_Settings_For_Memory_Usage_With_Large_Backups Yes, it is, but you have to restart it from the unRAID webui. I figured out why CrashPlan keeps restarting - it needs more memory. I think the 4.8 upgrade wiped out our previous adjustments to run.conf This article explains the problem. Unfortunately the article no longer contains the details on how to solve it, it just says to contact support. So the fix is: shutdown the CP docker SSH to your server type: cd /mnt/cache/appdata/CrashPlan/bin (or wherever your CP appdata files are) type: nano run.conf on the line that starts with SRV_JAVA_OPTS, change "-Xmx1024m" to "-Xmx2048m" press CTRL-O, CTRL-X type: exit start the CP docker This doubles the amount of RAM available to CP from 1 GB to 2 GB. The article above explains what CP recommends, you may need a different number. Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.