Enterprise Software

The Google Chrome startup mystery: Troubleshooting the Ghost in the Machine

Google Chrome has a handy self-updating feature, but this convenience can also require some detective work if things go wrong.

One of the many things I appreciate about Google Chrome is that it has the ability to update itself silently in the background without prompting or alerting you. Let's face it, too many applications are programmed to check for updates on startup or badger you when new updates are available. Updating an application before you can use it is like remodeling your kitchen to make a sandwich. Furthermore, I doubt there's a single user out there who opens their browser saying to themselves: "You know, now is really a great time to be notified of a new version of this program and asked whether I want to download and install it. I wasn't getting online for anything urgent anyhow."

Letting the robot maid do the housekeeping in the background is what technology was invented for. Having the robot ask me "Mr. Jetson is it OK for me to take out the trash now?" is not what I define as modern convenience.

I started using Chrome several revisions ago. However, to be honest, I've never been precisely certain which version I've been on thanks to the auto-update feature. Chrome is now up to version 21.0.1180.89 at the time this article was posted and I've only had a single update problem - however, the one I did have was memorable.

My adventure started last month when my 32-bit Windows 7 workstation at the office - which is always running - suddenly wouldn't open Chrome, which I leave up 24x7. No errors were produced. I tried launching the program several times, but without results.

Sifting through the evidence

I checked the Windows Application log, but found no references to the problem. Not to snipe at the competition, but this is a familiar issue when using Firefox, which sometimes snags on startup and shows in the Task Manager, but doesn't actually load, thereby necessitating that the process be manually ended and the program relaunched. However, in this situation I checked Task Manager and found no instances at all of chrome.exe running. That struck me as very weird. Normally when a program can't start there's SOME sort of reason why.

Since Chrome wasn't running, I didn't feel rebooting would change anything unless something else was blocking it somehow. Antivirus? I checked it, found it fine, and toggled it off. No joy, as we say in IT. I immediately toggled my antivirus protection back on.

The next inclination was to remove and reinstall Chrome. After all, thanks to Chrome Sync (check out a great article by my colleague Andy Wolber on how this feature works), getting all my bookmarks, settings, extensions and other personalized details back onto a reinstalled Chrome would be a snap. But this approach doesn't really tell us what's happened under the hood amidst a problem like this, and I wanted a direct answer if I could get it. As in the case of rebooting, I resisted the urge, at least for the time being.

As a Windows administrator I was tempted to try running Chrome in compatibility mode, which can be activated for any application in Windows 7 by right-clicking the program shortcut, choosing Properties, and then clicking the Compatibility tab as shown in Figure A.

Figure A

Compatibility tab

In most cases it's fine to check "Run this program in compatibility mode for:" and then leave the selection at "Windows XP (Service Pack 3)." Fine, that is, if the application will run in Compatibility Mode. I tried it in this case just to see what would happen, knowing that it probably wouldn't fix the issue since there's no reason Chrome should need to run in this legacy context (my hope was to get a different error which might give me some insight). No change in the situation.

I decided to take a look at the Chrome program files for any clues. I right-clicked the Google Chrome shortcut, and chose Properties. (Figure B)

Figure B

Clicking the "Open File Location" button displayed the following files under C:\Program Files\Google\Chrome\Application. (Figure C)

Figure C


I double-clicked chrome.exe but as before nothing happened.

I thought perhaps the issue might be with the location of the Chrome installation. Chrome is usually installed under the user profile folder; e.g. C:\Users\(user account)\AppData\Local\Google\Chrome\Application. If you navigate to this location in Windows to check out the files, please keep in mind that Mother Microsoft has hidden away many of the critical folders under your user profile folder. You'll need to go to Control Panel, Folder Options, click the View tab (Figure D) and choose the option to "Show hidden files, folders and drives."

Figure D

Folder options

In my case, I did have a C:\Users\(user account)\AppData\Local\Google\Chrome\User Data folder, but it contained no program files. This folder is for user-specific information such as bookmarks, history and last tabs. I went back to the C:\Program Files\Google\Chrome\Application folder. It had been used by Chrome before the error occurred, so I didn't see why its location should be a problem now.

In viewing the contents of the Application folder, the debug.log file (Figure E) immediately leaped out at me. I opened it in Notepad.

Figure E

Debug.log file

Wow! What was going on?

My "Aha" moment

Well, the issue was a lot plainer than the above screenshot might suggest. The line "Failed to load Chrome DLL from C:\Program Files\Google\Chrome\Application\21.0.1180.60\chrome.dll" told me that Chrome was trying to open from the C:\Program Files\Google\Chrome\Application\21.0.1180.60 folder, but couldn't do so.

I opened the 21.0.1180.60 folder and immediately the solution presented itself. It had only a handful of files as compared to the other folder, 21.0.1180.75 - and no sign of the chrome.dll file the error log reported missing. I copied all the files and subfolders from the 21.0.1180.75 folder into the 21.0.1180.60 folder and then Chrome started up. All my bookmarks and settings worked fine.

So, what happened?

It seems a problem occurred when Chrome attempted to update from one version to another. The update failed which left Chrome in an unusable state, until I copied the newer files to the old version folder. Because I leave Chrome up and running all the time (I use a free Windows utility called 4t Tray Minimizer which allows me to minimize running applications to the System Tray so they don't clutter up the Taskbar), that may have interfered with the update process and confused the program as a result.

It's possible to manually check for updates in Chrome and apply them if you (or one of your users) are having issues getting the latest version to install on its own. To do this, click the wrench icon (Figure F) in the upper right of the program:

Figure F

Wrench icon
Choose "About Google Chrome" (Figure G)

Figure G

About Google Chrome

In this case, my Chrome installation is up to date, but if an update were available the details would appear here. Of course, it's also possible to download the latest version of Chrome directly from Google.

What can you do to get further information?

Google offers a Chrome Help page with some basic details. There is also a Google Chrome Forum with many useful posts. If you're on Google+ you can also follow the Google Chrome page for some useful everyday tips as well as announcements about new releases.

Also read:


Scott Matteson is a senior systems administrator and freelance technical writer who also performs consulting work for small organizations. He resides in the Greater Boston area with his wife and three children.


Well, your Google Chrome data files are corrupt. Delete all app data files so that Chrome will re-create them again the next time you open it.

1. Press Windows Key + R, or click Start All Programs | Accessories Run.

2. Type AppData in the textbox.

3. Click OK or press ENTER.

4. Open following folder:

C:\Users\{Enter User Name Here}\AppData\Local\Google\Chrome\User Data\Default

5. Delete all files in the Default folder.



Those of us who've been in the IT world for any sigficant amount of time have been nailed by an update, a patch, etc which did more damage than good. Off the top of my head, I can recall the McAfee pattern update which quarantined svchost.exe on XP systems, a Symantec update which totally disabled the system, and various Windows patches which were quickly replaced with out-of-band releases when they had disasterous side effects. I like to be sure that there's not any chatter about an update before I commit to installing it. And I'd bet that most of us who've 'been-there-done-that' set the "Remind me of an update" option wherever we can, rather than allow updates to proceed automatically. There's a very good and compelling reason to make that choice. That said, I support home users where I've set the automatic update option, knowing that most of the time it functions as needed with no user decision required. But for my own PC's - and expecially in a centrally managed work environment, I think it's a potentially dangerous option.


Sure auto updating at home is fine, but in a business environment it could be an issue. At home if Chrome doesn't work then you can take some time and deal with it. At work if 1500 users suddenly can't access an internal web app because of some change Google has made then it becomes a major impact to business operations. Bill


I think a lot of applications could and should update themselves, but users typically don't allow them. I think this goes back to a lack of trust that the applications won't destroy themselves or others. I recall a Symantec update that crashed our computer to the point that it required restaging. Though this is the extreme case, it really highlights the problem. One possible improvement is that an application could offer a path backwards. Greater trust could be earned if it offered a way to undo the damage that it might create. This wouldn't have helped the afore mentioned crash, but it might be a step in the right direction.


Rosie the maid robot always behaved beneficially, and never did any spying or tracking for gain, and seemed to 'love' the Jetson family. Uniblab (Unablab?), on the other hand, was a conniving, sneaky robot, who spied on Mr. Jetson, tricked him and his co-workers into doing things he should not have, recorded the whole thing, and later showed the information all to the boss, Mr. Spaceley, for his own gain. (strictly for research requirements: "http://www.youtube.com/watch?v=cHzVy1R3lJQ") Which one is Google or Chrome more akin to? "Get to work! (click) Get to work! (click) Everybody! Work-work-work-work! (click) Work-work! Mr. Editor, please choose proper robot!


Wow! Lovely article. Got me thinking "so that's what happened!". Anyways Chrome has been my default browser for over 5 yrs now and I've always used it with firefox as a substitute but Maxthon has replaced firefox as my alternate browser. It's pretty impressive but my google integration is pretty important so chrome will always come out on top.


I have the same problem and went through your procedure, step-by-step. I didn't have a debug.log file though, so I'm still stuck with a non-working chrome. I found a solution to put the term "-no-sandbox" in my desktop shortcut, but this gives me warning messages about security and stability each time I start the program. Since it's the only way I can make it work though, that's what I do. I'm also amazed you're the first person to cover this topic. There are many forums on Google about the problem and nothing but baffled techs with no answers replying from Google. They went so far as to recommend using another browser and stating we shouldn't blame Chrome because it's free. Wow...


Uninstalls without user having admin rights. What a huge pain. We have webapps customizted and only supported with IE. When things blow up and they are related to a unapproved google chrome install. Walk away!

Mark W. Kaelin
Mark W. Kaelin

So far, I have been lucky and have had no problem with Chrome's automatic updates. How about you and your organization? Have you had to troubleshoot Chrome lately?


...Computing power, that is. I disagree with the original author's desire to have his programs automatically update themselves in the background without notifying him. If given the option, I always choose "notify me that an update is available" due to the fact that I typically work with older, less powerful computers at home and work. If I'm trying to do something, the last thing I need is for my processing power, hard drive access, and network bandwidth to get eaten up by a download and update that I have no control over. I don't like surrendering control of my system, particularly if it is unclear what is happening. For every one clueless user that will ignore any requests to update and has to have control of their compute power taken away "for their own good" there has to be a number of us who should not have our work disrupted at random and can be trusted to update as soon as it is practical to do so. Besides, not disrupting our work, if we know we have to update something, we can make sure that the update goes smoothly and can more easily recognize and deal with problems if they occur. The author mentioned turning off various utilities to ensure that Chrome would install correctly. If he knew when Chrome was going to install, he could have pro-actively turned the utilities off in advance and possibly avoided the issue entirely.


but then, I can't stand chromium's user interface, so I don't use it. :p

Editor's Picks