ReportCrash High CPU & How to Disable reportcrash in Mac OSX
For a while now, all of my MacBooks have run extremely hot and the fans have gone nuts. While troubleshooting the issue, Activity Monitor showed that an app named reportcrash has run very high on the CPU and has killed my battery life. Force quitting the app didn’t help as it would start right back up in a few seconds and climb back to 80-100% usage of the CPU.
What is CrashReporter?
CrashReporter runs in any time an application crashes and it’s designed to saves the application state to aid developers in working out why the app crashed. Basically a process is launching, crashing (and invoking CrashReporter) and then re-launching, repeating this cycle never ending.
How to Identify What’s Crashing
To show which process is triggering this cycle and stop it, CrashReporter is pretty verbose in its logging which makes finding the problem app somewhat easier. Open up the console.app (/Applications/Utilities/Console.app) and look towards the end of your system.log to see what app is crashing.
Unfortunately for me, the problem is a driver by some company called EFI and getting the latest drivers didn’t resolve the issue. The next obvious solution was to disable reportcrash.
How to Disable ReportCrash
Fire up terminal and run the following commands to disable reportcrash:
launchctl unload -w /System/Library/LaunchAgents/com.apple.ReportCrash.plist sudo launchctl unload -w /System/Library/LaunchDaemons/com.apple.ReportCrash.Root.plist
How to Enable ReportCrash
If you need to reenable crash report, run the following commands in terminal:
launchctl load -w /System/Library/LaunchAgents/com.apple.ReportCrash.plist sudo launchctl load -w /System/Library/LaunchDaemons/com.apple.ReportCrash.Root.plist
How to repair permissions on Linux Apache /var/www/html folder
I recently ran into an issue where I could no longer FTP files to my Linux droplet when multiple users were uploading to the server. The server kept the user as the owner despite me adding them to the www-data group. This fix comes from my ex-boss, James Tomasino who was kind enough to provide some help since I was stumped. I’m sharing in case anyone else runs into the same issue and finds it useful.
You’ll need to fire up terminal, ssh to the server, and then execute these commands:
cd /var/www sudo chmod 775 html sudo chgrp www-data html sudo chmod g+s html
+s makes permissions sticky so that all files will inherit from the parent directory. This was the setting I was missing.
Open up /etc/ssh/sshd_config. I use nano so:
sudo nano /etc/ssh/sshd_config
Hit CTRL+W and look for “subsystem” which is typically located near the bottom of the file. Change
subsystem sftp /usr/lib/openssh/sftp-server
to
subsystem sftp /usr/lib/openssh/sftp-server -u 0002
If you already have files in the HTML folder, you’ll want to run these commands to reset the permissions:
cd /var/www/
sudo chgrp -R www-data html
find . -type f -exec chmod 664 {} \;
find . -type d -exec chmod 775 {} \;
Regex to parse date formats when unknown
I’m working on an app that will parse different file sources and aggregate it for a report. Of course, each format has a different date format and trying to parse it all has proved to be a pain. I wrote a regex that’ll parse out just about every datetime format I’ve run into that I am sharing in case someone else finds it useful. I’ve put it on a Github gist along with a sample of the various dates I’ve tested it against and confirmed to work. If you find a format not covered by the regex, post a comment and I’ll update the gist.
Just a note that I haven’t finished parsing the timestamp (e.g. 1997-07-16T19:20:30+01:00) format. The date portion does get extracted correctly so I left it in.
(98)Address already in use: AH00072: make_sock: could not bind to address [::]:80 – error
I recently got handed a new Ubuntu droplet at work to setup and work on. I was going through the typical configuration to lock down the server and go to installing Apache when I suddenly ran into the error
(98)Address already in use: AH00072: make_sock: could not bind to address [::]:80 – error
I was a little surprised considering I had only run 4 commands total on a brand new installation. If you’re getting this error like me, you’ll need to find out what’s being bound to port 80 and then stop it. Use this command to get a list of anything that’s running on port 80.
sudo lsof -i:80
For me, it turned out nginx was running and bound to port 80. If you have the same problem, run this command to stop it:
sudo service nginx stop
Now you should be able to start apache.
How to Copy and Paste into the Digital Ocean VNC Console
I recently signed up for Digital Ocean to test droplets for my development and linux management testing. I was securing the server and setup fail2ban, disabled the root login, and created a user account but forgot to change the user name from root to the user name I selected in my SSH client profile. I ended up locking myself out of the server and had to resort to logging into the admin panel on Digital Ocean’s website and use their browser embedded VNC client to restore my access. I quickly discovered that one of the downsides of using this panel is that you can’t copy and paste commands. Luckily there’s a workaround and you can use the browser console to do sendkeys through Javascript.
Bring up the console in the browser developer tools. Here’s a cheat sheet for keyboard shortcuts:
| Browser | Description | Windows | Mac |
|---|---|---|---|
| Chrome | Open Developer Tools and bring focus to the console | Ctrl + Shift + J | Cmd + Opt + J |
| Firefox | Open Console | Ctrl + Shift + K | Cmd + Opt + K |
| Internet Explorer | After hitting F12, you have to click the console tab. There’s no direct shortcut to the console tab. | F12 | N/A |
!function(){function t(){window.rfb.sendKey(e.shift().charCodeAt()),e.length>0&&setTimeout(t,10)}var e=prompt("Enter text to be sent to console").split("");t()}();
Update 9-21-16
Ruden and Sebastiaan’s pointed out there was a bug in the code above. This updated snippet has support for characters when using shift characters like !@#$%^&*()_+
!function(){function t(){function n(t,e){s=s.concat(RFB.messages.keyEvent(t,e))}var o=e.shift(),s=[],i=o.charCodeAt(),c=-1!=='!@#$%^&*()_+{}:"<>?~|'.indexOf(o),r=XK_Shift_L;c&&n(r,1),n(i,1),n(i,0),c&&n(r,0),rfb._sock.send(s),e.length>0&&setTimeout(t,10)}var e=prompt("Enter text to be sent to console").split("");t()}();You’ll receive a dialog prompt to enter the copy to paste in. Please note that you need to click on the VNC console and hit enter to execute the command.
If you need to enter more than one command, just hit the up arrow on your keyboard in the console to get the script again for easy reuse.
Sitecore 8 Install Error: “The name you entered is not unique.”
I was working on my business partner’s computer to help him fix errors with his Sitecore instance for his certification class. He experienced multiple issues with the Sitecore 8 installer on Windows 8 and we went the manual route of deleting Sitecore to install clean and start from scratch. In the process, we:
- Deleted the IIS website instance in the IIS Manager
- Deleted the databases used by the instance
- Deleted the root folder in the filesystem
After launching the installer again, it kept returning the error “The name you entered is not unique.” when naming the Sitecore instance to one previously used. After checking the IIS metabase and a few other typical locations, couldn’t figure out where the Sitecore installer was finding the name since we deleted everything manually. It turns out that the Sitecore executable installer creates a registry entry under HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Sitecore CMS which is used to display information in Windows Programs and Features. This registry key is also checked during the installation process and if a child key with the same name as the one being installed is found, the above error message is shown.
To fix this:
- Start -> Run (or Windows Key + R)
- Type regedit
- Navigate to HKEY_LOCAL_MACHINE\SOFTWARE\Wow6432Node\Sitecore CMS.
- Delete the registry keys containing the information about the previously installed instance.
NOTE: You’ll need to relaunch the installer as it apparently caches the key information.
How to fix “hacked by Moroccanwolf” WordPress site
A client of mine had their wordpress hacked and when you would load the site, it would simply display a message that said “hacked by Moroccanwolf”. I did some digging and luckily it wasn’t a major hack and they didn’t mess with the posts or other settings as a lot of the hacks do.
Quick Fix
- To fix it, you’ll need to connect to your database using an editor of some sort, such as PHPmyadmin.
- Once logged in, expand your database on the left.
- Click on ‘wp_options’.
- On the top right, click the ‘Browse’ tab.
- Look for ‘widget_text’ in the option_name field. (For my client, it was at row 90). You should see something similar to this:
<script>document.documentElement.innerHTML = unescape(''%48%61%63%6b%65%64%20%62%79%20%4d%6f%72%6f%63%63%61%6e%77%6f%6c%66%20%26%26%20%61%62%64%65%6c%6c%61%68%20%45%6c%6d%61%67%68%72%69%62%69'');</script>
Delete the entire tag and this should restore your website. Now remember to change your logins and update wordpress.
How I Figured it Out
Most of the hacks I’ve seen are done through injecting javascript into the database somewhere which either force a redirect or something along those lines. Here’s the steps I followed to find it:
- To fix it, you’ll need to connect to your database using an editor of some sort, such as PHPmyadmin.
- Once logged in, click on your database on the left. Ensure you’re on the database and not a table.
- Click Export.
- Leave it to quick and click go.
- You should now see a textbox with a mess of SQL commands.
- Copy and paste into your favorite editor,
- Search for <script> and you should find something that doesn’t belong. In this instance, that was the only thing I found of note.
- You’ll want to scan the rest of the database for things that don’t belong. Additionally, you’ll want to replace all the wordpress files and confirm no .htaccess files were created that give hackers write access.
When updating Office 2011 for Mac, you are prompted that you must close “Microsoft Database Daemon” and “SyncServicesAgent”
If you’re trying to update your install of Microsoft Office 2011 for Mac, you may get the following message:
“These applications must be closed before the software can be installed:
Microsoft Database Daemon
SyncServicesAgent
Close these applications and try again”
If you try to kill the apps using activity monitor, you’ll find they keep getting restarted. The only way I found that worked to stop them is by following the instructions below.
- Open a terminal window and enter the following command:
launchctl unload ~/Library/LaunchAgents/com.microsoft.LaunchAgent.SyncServicesAgent.plist
This will unload the SyncServicesAgent. Unfortunately the Daemon kept running for me after the latest update. To stop that, do the following:
- Open activity monitor
- Scroll down until you find Microsoft Database Daemon in the list of active applications.
- Click on Microsoft Database Daemon and click on the Quit Process button in the toolbar
Go back to the install utility and you should now be able to proceed with the install. Once the install is finished, you’ll want to restart the Sync Services Agent, so do the following:
- Go back to the terminal window and enter this command:
launchctl load ~/Library/LaunchAgents/com.microsoft.LaunchAgent.SyncServicesAgent.plist
Now you should be all set with updates and can close terminal.
Steam is having trouble connecting to the Steam servers.
If you use Steam and suddenly find yourself unable to connect receiving the error “Steam is having trouble connecting to the Steam servers.”, check and see if you have Peerblock installed and running. I wasted a lot of time trying to figure out why the connections wouldn’t go through and finally went through apps running in the background one by one, until I closed Peerblock and discovered Steam suddenly started working. If you would like to continue to use Steam while Peerblock is open, follow these instructions for a workaround:
1. Start PeerBlock
2. Click List Manager
3. Click Add
4. Put something in the description like steam-allow
5. click add url and paste this: “http://list.iblocklist.com/?list=steam” (no quotes)
6. Choose type: allow
7. Click Ok then close the List Manager.
This will download a list that adds exceptions for Steam and should let you connect with no issues.
Visual Studio 2013 package did not load correctly
Today I loaded up Visual Studio to continue work on a project and created by a random bunch of errors. All were complaints that various packages could not be loaded. The one I captured was:
[alert color=”gray”]The ‘JavascriptWebExtensionsPackage’ did not load correctly.’ The problem may have been caused by a configuration change or by the installation of another extension.[/alert]
I had just installed SyncFusion’s Essential Studio which apparently screwed something up. I tried doing a repair on Visual Studio which unfortunately did not work.
I was able to fix it by following these steps:
- Close Visual Studio Open the *Users*\AppData\Local\Microsoft\VisualStudio\12.0\ folder
- Rename the ComponentModelCache folder
- Restart Visual Studio.
Visual Studio should now rebuild the cache and no longer display the error messages.