xusia

Members
  • Posts

    10
  • Joined

  • Last visited

Posts posted by xusia

  1. Did you try and stop and start the array after the repair?

     

    When I had to run reiserfsck to do a repair years ago - immediately after the repair nothing showed up. But when the drive was re-mounted, all of the files were there.

     

    I'll also note that the practice of removing a drive, starting the array, re-installing the same drive, and then doing a rebuild is inherently dangerous. When a cable drops the drive stops being updated, but all of the files that are on the disk at the time is was dropped are still there. Losing a few new files is usually not a huge issue compared to losing the entire disk. When you do a rebuild on top of that disk, you are putting all of your eggs in one basket. It is a far better approach to rebuild onto a different disk. Then, if anything goes wrong with the rebuild, you still have the original disk with all or most of your data as a backup.

     

    What I like to do is remove the disk and confirm that the disk is being properly simulated (some call it emulated, same thing). Only if the simulation looks good would I do a rebuild. If someone wants to do a rebuild after confirming the simulation looks good, I consider it a relatively small risk for an experienced user. But it is a risk and they should know that rebuilding onto a different disk is the safer option. For relatively new users, I would recommend the rebuild onto a separate disk. Running a second recovery exercise after a rebuild should only be attempted after a lot of diagnosis and vetting. Because it is so easy to make things worse - and no way to undo.

     

    Thanks for the info. If I can't get the data back I will remember this for next time. It makes perfect sense and I'm kicking myself for not having done it this way in the first place.

  2. Hi folks, I've searched the forum and haven't found this specific problem. So here goes. Hope you can help.

     

    All was well with my array when a drive dropped out showing disabled. It turned out to be the usual culprit - crook SATA cable. Fixed it up, and got the disk showing again. I stopped the array, removed the redball disk. Started the array, stopped the array, added the disk back in. Started the array and data rebuild proceeded as per normal. However, when data rebuild was complete, the disk showed over 1 TB of data in use, but when I accessed the disk to view the files - there was nothing there. Filesystem is xfs - I tried xfs_repair - it looked like it did something, but still no data. Syslog is attached. You may see by the attached syslog that a data rebuild is currently in progress - this is because I repeated the procedure in the hope that the data rebuild would succeed a second time. Anybody got any ideas?

    xusia-syslog-20150830-1706.zip

  3. I have done a bit of study and worked it all out myself. Here is what I found out:

     

    1. You cannot expose another host volume to a running container. They all have to be there in the command line on the original "docker run" command.

     

    2. If you update say, Sickbeard inside the running container, then you can save the changes permanently using docker commit, but in order to guarantee that you never lose those changes, you then have to use "docker push" to push them up to your own repository. This involves creating an account on the docker hub, and then pushing the image changes to the hub under your own username into a new image after using docker commit to save the container changes. Then if you ever have to delete the container and recreate it, you refer to your new repository when using docker run, instead of "eschultz/docker-sickbeard" - the same will work if you update any of the other apps. I can post a more detailed set of instructions if anyone wants me to.

  4. Hi all, I've got docker running now using the excellent quick start guide on the forums. Here's what happened to me yesterday, followed by a couple of queries:

    I got sickbeard running, and it advised me that a newer version was available, so I updated it, restarted the container and was up to the newest version. Then I realised that I had to expose a couple more host directories to the container as volumes. I couldn't find a way of adding another volume without deleting and re-creating the container. Consequently once I re-created the container with the extra volumes I wanted, sickbeard was back to the version in Eric Schultz's image, and the sickbeard database wouldn't open because it had been upgraded to the latest version. So I deleted the sickbeard database files, and was able to restart sickbeard OK then. OK, on to the questions:

     

    1. Can you expose an additional host volume to a container without deleting and recreating the container?

    2. If I update sickbeard inside its container, can I use "docker commit" to commit the changes so sickbeard stays upgraded permanently?

    3. If docker commit can be used, would I then have to launch a local copy of the upgraded container?

    4. If I do the upgrades, will everything then fall apart if Eric releases any updates to his images?

     

    Thanks.

  5. Hi all. Since updating to V6 Beta 10a I have noted a frustrating issue with the WebGUI. I use the latest version of Google Chrome, and address my server by its IP rather than a DNS name ie http://192.168.1.129 (that's the static IP I have allocated to the server). The interface seems to be intermittently unavailable. When I first log in, I get asked for credentials. After entering the root creds, when I press the login button, the request for credentials reappears. Sometimes I have to press login up to five or ten times, but eventually get to the main menu. Then if I want to go to the Settings menu or any other menu, I often get a "not found" message. Repeated browser refresh eventually gets me into the page. If I leave the page open and come back to it say a half hour later, I hit refresh and get asked for credentials again (as if the session expired - which didn't seem to happen in previous versions). Have I got a defect here, or have I done something silly or missed setting a preference somewhere? syslog attached. Let me know if you need to know hardware specs.

    unRAID_syslog_29-10-2014.txt