GNOME Bugzilla – Bug 762565
Content indexing spamming syslog and logging personal information (emails etc) - there should be a way to disable file contents indexing
Last modified: 2016-05-08 19:31:54 UTC
I have noticed some problems similar to this bug report in Tracker's logging to syslog: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=754907 So what I am noticing is that it is spamming my syslog with loads and loads of stuff similar to the bug report I linked to above. But not only is it spamming contents there, it is also putting the contents and details of my private emails etc, for instance I have OSSEC HIDS running and it keeps on reporting to me of these sorts of things: " OSSEC HIDS Notification. 2015 Jul 22 21:08:46 Received From: Weaponised-Towel->/var/log/syslog Rule: 1003 fired (level 13) -> "Non standard syslog message (size too large)." Portion of the log(s): Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nie:description "Beginning Android Games, Second Edition offers everything you need to join the ranks of successful Android game developers, including Android tablet game app development considerations. You'll start with game design fundamentals and programming basics, and then progress toward creating your own basic game engine and playable game apps that work on Android and earlier version compliant smartphones and now tablets. This will give you everything you need to branch out and write your own Android games. The potential user base and the wide array of available high-performance devices makes Android an attractive target for aspiring game developers. Do you have an awesome idea for the next break-through mobile gaming title? Beginning Android Games will help you kick-start your project. This book will guide you through the process of making several example game apps using APIs available in new Android SDK and earlier SDK releases for Android smartphones and tablets: The fundamentals of game development and design suitable for Android smartphones and tablets The Android platform basics to apply those fundamentals in the context of making a game, inc --END OF NOTIFICATION " And then when I checked my syslog for that time period these are the sorts of things that I saw: " Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: INSERT { Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: GRAPH <urn:uuid:472ed0cc-40ff-4e37-9c0c-062d78656540> { Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: <urn:uuid:edaa188c-36cb-b2d3-0cb3-97c950984b51> nie:dataSource <http://www.tracker-project.org/ontologies/tracker#extractor-data-source> . Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: <urn:uuid:edaa188c-36cb-b2d3-0cb3-97c950984b51> a nfo:EBook ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nco:creator [ a nco:PersonContact ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nco:fullname "Zechner, Mario & Green, Robert" ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nco:nameFamily "Zechner" ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nco:nameGiven "Mario" ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nco:nameAdditional "& Green, Robert" ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nco:creator <urn:artist:Zechner,%20Mario%20&%20Green,%20Robert>] ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nco:creator [ a nco:PersonContact ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nco:fullname "Robert Green" ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nco:nameFamily "Green" ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nco:nameGiven "Robert" ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nco:creator <urn:artist:Robert%20Green>] ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nie:title "Beginning Android Games" ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nco:publisher [ a nco:Contact ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nco:fullname "Apress"] ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nie:identifier "URN:ISBN:978-1-4302-4677-0" ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nie:identifier "9781430246770" ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nie:description "<p class=\"description\">Beginning Android Games, Second Edition offers everything you need to join the ranks of successful Android game developers, including Android tablet game app development considerations. You'll start with game design fundamentals and programming basics, and then progress toward creating your own basic game engine and playable game apps that work on Android and earlier version compliant smartphones and now tablets. This will give you everything you need to branch out and write your own Android games. The potential user base and the wide array of available high-performance devices makes Android an attractive target for aspiring game developers. Do you have an awesome idea for the next break-through mobile gaming title? Beginning Android Games will help you kick-start your project. This book will guide you through the process of making several example game apps using APIs available in new Android SDK and earlier SDK releases for Android smartphones and tablets: The fundamentals of game development and design suitable for Android smartphones and tablets The Android platform basics to apply those fundamentals in the context of making a game, including new File Manager system and better battery life management The design of 2D and 3D games and their successful implementation on the Android platform This book lets developers see and use some Android SDK Jelly Bean; however, this book is structured so that app developers can use earlier Android SDK releases. This book is backward compatible like the Android SDK. What you’ll learn How to set up/use the development tools for creating your first Android game app The fundamentals of game programming in the context of the Android platform How to use the Android's APIs for graphics (Canvas, OpenGL ES 1.0/1.1), audio, and user input to reflect those fundamentals How to develop two 2D games from scratch, based on Canvas API and OpenGL ES How to create a full-featured 3D game How to publish your games, get crash reports, an Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: d support your users How to complete your own playable 2D OpenGL games Who this book is for This book is for people with a basic knowledge of Java who want to write games on the Android platform. It also offers information for experienced game developers about the pitfalls and peculiarities of the platform.</p>" ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: nie:language "en" ; Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: (tracker-extract:1740): Tracker-WARNING **: Task 13, error: Subject `urn:uuid:7f45658e-fe34-a168-af48-c6bdb3901b6c' is not in domain `nfo:Image' of property `nfo:orientation' Jul 22 21:08:44 Weaponised-Towel gnome-session[1509]: (tracker-extract:1740): Tracker-WARNING **: Sparql update was: " And after some further research I have come to understand that this is mostly happening because there is no way to disable the feature which indexes file contents. So it would be really good if a feature like that could be added, or at least for you to be able to control the logging levels. Another reason why it would be really good if there was at least an option to turn off content indexing and just have file name indexing is that it can sometimes be really hard to find anything if say you know the file is called "account", however you have very many files on your machine with that word in them. So you will get so many results it may be hard to find what you are actually looking for. Some users may also want to disable content indexing because of performance issues. When I initially filed the report here (https://bugs.launchpad.net/ubuntu/+source/tracker/+bug/1477489) I was running Ubuntu GNOME 15.04, but now I am running Ubuntu GNOME 15.10 with GNOME 3.18 and the issue is still present.
(In reply to cooks.go.hungry from comment #0) > I have noticed some problems similar to this bug report in Tracker's logging > to syslog: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=754907 > > So what I am noticing is that it is spamming my syslog with loads and loads > of stuff similar to the bug report I linked to above. But not only is it > spamming contents there, it is also putting the contents and details of my > private emails etc, for instance I have OSSEC HIDS running and it keeps on > reporting to me of these sorts of things: <snip> You will be glad to know logging got less verbose in master/1.9. Also, there's good chances that the bug producing the syslog excerpt you bothered to paste was fixed as of 1.8.0. I'm saying "good chances" because the snippet is partial, you cut off relevant information, so I can't tell for sure. > > And after some further research I have come to understand that this is > mostly happening because there is no way to disable the feature which > indexes file contents. So it would be really good if a feature like that > could be added, or at least for you to be able to control the logging levels. There is such way to control logging level, you can manipulate the /org/freedesktop/tracker/extract/verbosity setting with dconf-editor. But unless you modified that yourself, you're already at the minimum logging level. I'm not sure if you realized that those warnings are pointing out bugs that need fixing. We obviously don't have access to files triggering this specific behavior (otherwise the bug would be fixed and there'd be no warnings), so this logging is our only chance to improving Tracker thanks to external input. > > Another reason why it would be really good if there was at least an option > to turn off content indexing and just have file name indexing is that it can > sometimes be really hard to find anything if say you know the file is called > "account", however you have very many files on your machine with that word > in them. So you will get so many results it may be hard to find what you are > actually looking for. > Some users may also want to disable content indexing because of performance > issues. There's already enough settings to control what, when and how things are indexed. A global on/off toggle is one that shall not be added, disabling tracker will leave an indefinite number of applications depending on it crippled or plain inoperative. So, sorry, but the requested change won't happen, I'm closing this bug.