Half a Page of Scribbled Linux

These instructions are using ROM Manager 5.5.3.0. The radio upgrade process has gotten much easier since I last did it.

From your computer, download the radio zip file (Toro radios are available from xda-developers) and check its MD5Sum. Put the radio zip file in Dropbox. (If you don’t have Dropbox, you can email it to yourself or connect USB and transfer it that way.)

On the phone, open the zip file in Dropbox (or a file manager). Depending on what apps you have installed, you’ll have the option of opening it with ROM Pre-Installation or ROM Pre-Installation will start automatically. For radios, there’s no need to change the defaults (don’t back up existing ROM and don’t wipe cache and data). Select “Reboot and Install”. That’s it!

On my Samsung Galaxy Nexus, I have CyanogenMod 10.1, which is based on Android 4.2. In previous versions of Android, I would copy the ClockworkMod backups from /data/media/clockworkmod/backup to my PC. In Android 4.2, connecting with adb shell won’t let you look at the data directory. To fix this, you need to set root permission in Settings > Developer Options > Root access. (If you don’t see Developer Options, it’s because it’s hidden by default in 4.2 and up. Go to Settings > About phone and tap Build number seven times to enable it.) Once in Root Access, select either ADB only or Apps and ADB. Then run adb root before adb shell. Now you’ll have permissions to see the /data directory.

The ClockworkMod backups are still in the same location, so you can get them with adb pull /data/media/clockworkmod/backup.

http://www.webupd8.org/2012/01/install-oracle-java-jdk-7-in-ubuntu-via.html

I have a show on my old Series 2 TiVo that I want to watch on a Linux machine. The TiVo Desktop software only runs on Windows. I downloaded it (TiVoDesktop2.8.3.exe and tivo-desktop-patch-setup.exe) to my tivo folder on the NAS. Next, I started a Windows XP session on VirtualBox. In Windows Explorer, I entered the NAS’s address (\\ds411p2) in the address bar and logged in. From there, I could get to the tivo folder. I ran TiVoDesktop2.8.3.exe and set up the software, then ran the patch.
TiVo Desktop cannot find my DVR, though the DVR is accessible via its IP address.

As a sanity check, I rebooted into a native copy of Vista. TiVo desktop ran just fine. This doesn’t satisfy the requirements of the title of this post, but Galleon should. Isn’t it funny that I disabled it only this week?

While reading a PDF, I noticed GKrellM was showing disk activity. I was curious to know what was causing that. Googling “top for disk access” turned up IoTop. Conveniently, this is in the Ubuntu repositories. I installed it with Syntaptic and ran it with sudo iotop. To make it easier to see what’s continually using the disk, press “a” to switch the display to accumulated access.

The disk access was from Galleon, the TiVo media server. I didn’t remember having it running and wasn’t using it, so I turned it off with sudo service galleon stop and disabled it with sudo update-rc.d -f galleon remove, which is the Ubuntu equivalent of chkconfig.

The disk usage went down, but was still going. Next, I saw activity with rsyslogd. In /var/log, the most recently written logs were user.log, syslog and debug. Looking at these showed the rotten Lexmark printer driver at work. It periodically scans for scanners and fills these logs up with debug messages. The last time I called Lexmark about it, they said they couldn’t do anything about it. That was about a year and a half ago. Looks like it’s time to look into that again. That’s another thing for “the list”.

There’s another thing accessing disk, and that’s jdb2, which is part of the journaling file system. I found a post about HAL causing this. This will also go on the list.

Note, this isn’t the most direct way to update, it’s only the path I followed.

1. Start ROM Manager
2. Boot into recovery
3. Select Backup and Restore, then Backup
4. Copy the backup to your computer
5. Download a CyanogenMod 10.1 M release (I used M3)
6. Run adb push cm-10.1-20130411-EXPERIMENTAL-toro-M3.zip /sdcard/.
7. On the phone, select “wipe cache”
8. On the phone, select “install zip from sdcard”, “choose zip from sdcard”, select update.zip above.
9. Reboot. Don’t freak out when your WiFi or Mobile connection indicators don’t turn blue. Until the Google apps are updated below, your phone won’t be able to connect to the Google servers.
10. Check your mobile network setting (I didn’t need to change this.)
11. Fix the clock
12. Download the updated Google applications
13. Reboot into recovery
14. Run adb push gapps-jb-20130301-signed.zip /sdcard/0/. (Note the path is different for Android 4.2!)
15. Load with “install zip from sdcard” as above
16. Wipe the cache
17. Reboot. Until you reinstall Gmail and Sound Search, you’ll get occasional messages they’ve crashed. That’s OK.
18. Sign in to Google. If you have 2-step authentication enabled, enter an application specific password.
19. Using Google Play, uninstall and reinstall Gmail and Sound Search. Although I didn’t see this documented, I also had to reinstall Amazon Mobile, Chrome, Google Play Music, Google Voice, ROM Manager and YouTube. I ended up going through the apps list one by one to check everything.

The whole process took about three hours including the usual household interruptions and a 45 minute wait to download GApps.

As I mentioned, I took about 1,200 photos in Israel. Since these are family vacation photographs, I collaborate with my wife to decide which ones get posted to Gallery on Family Tidings. 1,200 photographs is a lot of decisions. Luckily, Gallery has some tools to help.

First, I create an album for the pictures and make it hidden. Then I take all the candidate photos and upload them to the album. At this point, I can log in and rate them from one to five stars. My wife does the same from her account. After this is done, I keep the ones that have at least three stars. To do this, create an album called “extras” under the subject album. Mark it hidden on the General tab and exclude it from display in the image block on the Album tab. Edit the subject album and change the sort order to “rating”, descending. Back in the album, find the first photograph rated two or lower and from its menu, select “Move photo”. Select all the photographs after that, but make sure you don’t include the new “extras” album. For the Destination, choose the “extras” album under the subject album and select “move”. Select the subject album, choose Edit Album and set the sort order back to “default” and remove the hidden attribute.

I just got back from Israel and have about 1,200 pictures with an EST (UTC-5) timestamp that should have an IST (UTC+2) timestamp because I forgot to change my camera’s time. There’s a utility called ExifTool that can read and write the Exif (Exchangeable Image File) format embedded in the Canon Raw images with CR2 extensions. It’s available on Ubuntu in the libimage-exiftool-perl package. To change the dates, run
exiftool -AllDates+=HH:MM FILES.CR2.
In my case, my camera’s clock was 16 minutes fast and I had to shift the time ahead by seven hours, so I used
exiftool -AllDates+=6:44 *.CR2

This renamed each file to file.CR2_original and replaced the original with the updated Exif data. Unfortunately, the new file has the current modification time. I should have used the -P option. To fix this, I ran (in Bash):
for FILE in *.CR2; do touch -r ${FILE}_original ${FILE}; done

Alternately, I could have had Exiftool not rename the originals, but put the updated copies in another directory:
exiftool -P -AllDates+=6:44 -o ../from-camera-with-corrected-times *.CR2

After having the DS411+II for over a year, I updated the DSM version. It was version 3.x and now it’s 4.x. The new version has better support for packages, sort of like the Google Play store. This is really nice, but I can’t use it because I’m in a limbo state right now, as I have a mix of old third party packages and source installs. These are all related to Sickbeard and SABnzbd. SAB is recognized by the package manager as installed, but Sickbeard isn’t because it’s a source install.
There’s a new third party package repository called Synocommunity that contains these packages. I added that repository, but don’t want to upgrade Sickbeard until I’m confident I can migrate my configuration without hosing it.
The version I have of Sickbeard is the alpha master from Github. For some months, it has had a bug that prevents it from updating. The first thing I need to do is fix that and get Sickbeard up to a recent version.
Step one: Find the Sickbeard installation. The Sickbeard installation page has a link to a Synology forum post that links to a manual Synology installation post.
I SSH’d into the NAS as admin and ran find / -type d | grep -i sickbeard. The source tree is in /volume1/opt/share/sickbeard. I changed to that directory and ran git status. This showed I have no changes in my local repository and should be able to update manually with git pull, so I did and I got “error: cannot open .git/FETCH_HEAD: Permission denied”. The file exists and is readable, but only writable by root. I logged back in as root and ran git pull, but git wasn’t found. I added /opt/bin to the PATH and ran the pull again. This time it connected to the remote repository, but failed because it couldn’t find tr, which is in /volume1/@appstore/baseutils/utils/bin. I added that to the PATH and reran. Now the pull worked. I restarted Sickbeard and it’s now fine and the version check works.

Boon is a dual boot desktop machine with an old, mostly unused copy of XP in one partition and a LVM of rotating Linux distributions, currently Ubuntu 10.04. This machine is kind of cranky and doesn’t like being left on all day. Sunday I left it on all day. When I came back to it that evening, it had locked up. I rebooted and got a kernel oops. This morning I tried diagnosing the problem. The current version of Linux failed to boot. The prior version also failed. And the rescue version failed. The kernel oops listed an IO error, so I suspected a problem with the suspend image stored in the swap partition. (This machine is normally suspended rather than powered off.)

I started SystemRescueCD and it was just fine. So no memory or CPU problems. I ran vgdisplay to see the volume groups. I was able to mount the root and home partitions and they looked OK. In the root partition, I looked in etc/fstab to see where the swap partition was. It was referenced by a UUID. To check, I tried mounting that UUID and confirmed that mount saw it as a swap partition. To get rid of the suspend image, I ran mkswap on /dev/disk/by-uuid/.

After a reboot it’s back up!