Jumping on the Bandwagon

The Whine Percent

Wed, 23 May 2012 01:31:12 GMT

The Daily Mail is running snippets of communications from Wall Street bankers and executives complaining about how challenging it is to survive on their six-figure salaries. Andrew Schiff, the communications and marketing director at Euro Pacific Capital, said, “I’m crammed into 1,200 square feet. I don’t have a dishwasher. We do all our dishes by hand.” Alan Dlugash, a partner at accounting firm Marks Paneth & Shron LLP in Manhattan, said, “Could you imagine what it’s like to say, ‘I got three kids in private school, I have to think about pulling them out?’ How do you do that?” Dlugash added, “People who don’t have money don’t understand the stress.” - The Progressive May 2012

Cheapo Android Tablet DFP7005 FTW!

Thu, 03 May 2012 18:26:09 GMT

So I purchased an AGPTek DFP7005 Android tablet. The specs are as follows:
Model: DFP7005;
Android Version: 4.0.3 (ICS);
Kernel Version: 3.0.8+ shen@compiler #1;
Build Number: DFP7005_0.5.3-r167_20120322;
1GB Ram;
2GB Internal Flash;
External SDHDC slot;
LCD Resolution: 7" 800 x 480 pixels
Output Cable: HDMI & CVBS Output to HDTV;
Processor: A10 1.5GHz Multi Core
PMU (Power Management system ): AXP209
GPU ( 3D Graph Engine): Mali-400 GPU @350MHZ
Pixels: Front Camera 0.3MP

Excellent developer options, and I think it's already rooted. Part of the problem that I have is that many apps are "not compatible" with the device. For example, I cannot install "Google Maps" nor can I install the YouTube app. It's rather annoying, but I can deal with it at $125 for the whole package (note: the website versions of those apps work just fine). I really wanted to buy it in lieu of painting my self into a corner with a dedicated E-Reader (i.e.: kindle, or nook) that pretty much did one thing only. So far so good with this tablet. It fits into my jacket pocket, and battery life is great.

Now the only thing I need is the tun_tap.ko kernel module for this thing.

The most decadent sandwich

Sun, 11 Mar 2012 19:28:03 GMT

While watching the IT Crowd it came to me. A sandwich made of: turkey, provo, mayo, tons of fresh ground pepper between each layer, 1 hard boiled egg sliced, 1 fresh sliced garlic clove (as thin as you can make it, but not minced), pickled moroccan string beans (a lattice under the provo), multigrain bread, open faced into toaster oven at 450F for 15min. Note: Keep eggs and garlic on same side so they don't spill when putting it together after toasting - or put the garlic with the moroccan string beans *under the provo*.

Eclipse's MenuBar Sucks

Thu, 08 Mar 2012 14:44:38 GMT

I have to say it... and there. It's not the lack or plurality of items. It is the fact that when I click it and move the mouse the entire window becomes a floating window instead of a fully expanded window. While going for the time, I end up moving my IDE to the next desktop: very annoying. <EOF>

milter-greylist fails on Fedora 15 upgrade as socket "unsafe"

Wed, 22 Feb 2012 01:49:54 GMT

I was getting the following error from sendmail after upgrading from Fedora 14 to Fedora 15 (x86_64):

Milter (greylist): local socket name /var/run/milter-greylist/milter-greylist.sock unsafe
Milter (greylist): to error state

After a couple of hours, and much hunting on the internet, I realised that it was not a standard answer (i.e.: permissions, etc). Instead the daemon had not been started. Apparently sendmail has been moved to the new systemd (systemctl) and greylist was still using (only) upstart. So the daemon was not starting when sendmail would be kicked off. I created an sysv init script to kick it off and everything works.

Side note: Fedora please don't switch init systems *again*. If anyone wants to create a systemd init script for it I would love it! ;) I have to relearn the init process and am too busy at the moment.

Linux Durable (aka: last plank unchanged)

Thu, 09 Feb 2012 04:13:11 GMT

So the only thing remaining that had *any* contact with my first (IBM PC Compatible) computer is a video card that I still use as my primary machine. The graphics card: an ATI Experts-at-play card from Microcenter. First one that X11 supported a decent resolution on.
01:06.0 VGA compatible controller: ATI Technologies Inc 3D Rage Pro 215GP (rev 5c) (prog-if 00 [VGA controller]
)
        Subsystem: ATI Technologies Inc Rage Pro Turbo
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping+ SERR- FastB2B- DisINTx-
        Status: Cap- 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
        Latency: 32 (2000ns min), Cache Line Size: 32 bytes
        Interrupt: pin A routed to IRQ 5
        Region 0: Memory at fd000000 (32-bit, prefetchable) [size=16M]
        Region 1: I/O ports at ac00 [size=256]
        Region 2: Memory at feaff000 (32-bit, non-prefetchable) [size=4K]
        [virtual] Expansion ROM at fea00000 [disabled] [size=128K]
        Kernel modules: atyfb

Create and Add Memory on the Fly without Rebooting...

Wed, 01 Feb 2012 20:19:57 GMT

I had a long 3+ hours running process. OOM_Killer had already killed it before. This time I watched the amount of memory used rise to 300M shy of 2GB. In a panic, and not wanting to restart the VM to add more memory I just added swap:

dd if=/dev/zero of=extra_swap.img bs=1M count=1000
mkswap extra_swap.img
swapon extra_swap.img

Just add water! Instant 1GB of swap! Note, you would need the free disk space to begin with, but I suppose you can mount a CIFS or network share and place the file there. The only problem now is speed.

Using Firefox 3.6 with Thunderbird 8 ("Send Link...")

Wed, 25 Jan 2012 16:22:53 GMT

I need to use friefox 3.6 for applications here at work (Domino Web Access - iNotes) So even though I had everything defined correctly, the "Send Link.." wasn't working to send email via thunderbird. I realised in the stderr on firefox it would say:

/usr/lib/thunderbird/thunderbird-bin: /opt/firefox/libnss3.so: version `NSS_3.12.10' not found (required by /usr/lib/thunderbird/libxul.so)

So I guessed that this had something to do with the fact that I was still running a stock Firefox 3 and library incompatibilities with the newer thunderbird. Instead of downgrading tbird, I decided to edit the /usr/bin/thunderbird startup script. I first debugged it buy putting an "env|sort" right before it kicked off the app at the bottom. Found all references to /opt/firefox and overrode them, pointing to /usr/lib/thunderbird instead.

LD_PRELOAD=/usr/lib/thunderbird/libxul.so
DYLD_LIBRARY_PATH=/usr/lib/thunderbird/
SHLIB_PATH=/usr/lib/thunderbird/
LIBPATH=/usr/lib/thunderbird/
LIBRARY_PATH=/usr/lib/thunderbird/
LD_LIBRARY_PATH=/usr/lib/thunderbird/:/usr/lib/oracle/11.2/client/lib:/usr/local/lib
/usr/bin/thunderbird "$@"

Put that in a file called /usr/bin/thunderbird-firefox and made that the default email client.

Another Dooring - AKA reasons for dedicated bike lanes

Tue, 10 Jan 2012 13:50:46 GMT

I was walking to my building today from my bike and sure enough on 9th and Sansom, a bicyclist got doored. Three of us quick ran into the street to stop the traffic; she was still sprawled out. She wasn't wearing a helmet (and hit her head, as she said she was dizzy) and I think she might have dislocated her knee. I called the ambulance and waited until they got there. A neuro doctor came over and I helped him get her out of the street. Luckily he knew the right questions to ask and stuff; I was kind of freaking out in my own head. And just like a doctor, lectured her about wearing a helmet. I was like, "now's not the time." Apparently, she works in my building, so I took her bike into the racks and gave her my number just in case. This is exactly the reason why dedicated bike lanes are necessary in the city. Particularly bike lanes that are NOT adjacent to parked cars.

Upgrading to Fedora 15

Mon, 09 Jan 2012 14:36:27 GMT

I downloaded the DVD and ran the upgrade off of the dvd. Ran into issues with the RPM db where some packages conflicted when I tried to do a yum upgrade. So I had to upgrade those packages individually. After handling them individually, I could then do a full yum upgrade unattended. Some had to be completely removed (i.e.: tsclient, libpanelappletmm, etc.)
Second problem I ran into was ugly fonts. Basically TrueType Bytecode Interpreter (BCI)‎ patents have expired. So they enabled it in the freetype package that comes with the distro. This actually made the fonts look WORSE. So I had do download and install freetype-freeworld from the rpmfusion-free-updates repository. In addition, I had to create a file "/etc/fonts/conf.avail/99-autohinter-only.conf":

<?xml version="1.0"?>
<!DOCTYPE fontconfig SYSTEM "fonts.dtd">
<fontconfig>
  <match target="font">
    <edit name="autohint" mode="assign">
      <bool>true</bool>
    </edit>
  </match>
</fontconfig>

Then:
ln -s /etc/fonts/conf.avail/99-autohinter-only.conf /etc/fonts/conf.d

After a reboot, the fonts looked WAY better.

Portal Error Uploading Module

Tue, 29 Nov 2011 21:35:23 GMT

I was getting the following errors attempting to upload a portlet via the "install module" portlet.

com.ibm.wps.pe.mgr.exceptions.AppServerWarInstallException:
EJPPE0013E: The installation of Web Module from WAR file
/opt/IBM/WebSphere/wp_profile/PortalServer/deployed/Bookmarks.war did
not complete successfully when invoking the WebSphere Application Server
administration interface.

I resolved this by:
1. Stop the dmgr, nodeagent, and Portal servers
2. Clear out these directories (but leave the directories themselves intact):
<dmgr_profile_root>/temp
<dmgr_profile_root>/wstemp
<dmgr_profile_root>/tranlog
<wp_profile>/temp
<wp_profile>/wstemp
<wp_profile>/tranlog
3. Restart the servers.

IHS WAS and SSL Woes

Mon, 31 Oct 2011 16:01:17 GMT

IHS HTTP Config for SSL

To enable SSL for IHS, use iKeyMan to generate a new key database file in /opt/IBM/HTTPServer/conf directory
Create a key store password stash file in the same directory. This will be your KeyFile (i.e.: ihsserverkey.kdb)
Click on "New Self-Signed..." if that's what you would like to
use, or import a signed certificate for your CN (i.e.:
ihs01.example.com)
Make note of the certificate label. This will be your SSLServerCert (i.e.: ihs01)
Edit the httpd.conf file adding or replacing the SSL configuration section as such:

<code>
# Example SSL configuration which supports SSLv3 and TLSv1
# To enable this support:
#   1) Create a key database with ikeyman
#   2) Update the KeyFile directive below to point to that key database
#   3) Uncomment the directives up through the end of the example
#
LoadModule ibm_ssl_module modules/mod_ibm_ssl.so
Listen 443
<VirtualHost *:443>
SSLEnable
# label of key in kdb
SSLServerCert ihs01
SSLProtocolDisable SSLv2
</VirtualHost>
KeyFile /opt/IBM/HTTPServer/conf/ihsserverkey.kdb
SSLDisable
# End of example SSL configuration
</code>

The next part configures the plugin to talk via SSL with WAS. Note
that these steps are necessary particularly if you're getting the
following error in the Plugins/logs/webserverName1/http_plugin.log:
ERROR: lib_stream: openStream: Failed in r_gsk_secure_soc_init: GSK_ERROR_BAD_CERT(gsk rc = 414) PARTNER CERTIFICATE DN=CN=XXX,OU=XXXCell01,OU=XXX,O=IBM,C=US, Serial=12:3d:40:f1:dc:XX:XX:XX

Go to your deployment manager for WAS
Navigate to:
- Network Deployment: SSL certificate and key management > Key stores and certificates > CellDefaultTrustStore > Signer certificates
- Stand Alone Deployment: SSL certificate and key management > Key stores and certificates > NodeDefaultTrustStore > Signer certificates
Click the checkbox next to the CN=YOURHOSTNAME, OU=Root certificate and click extract.
Copy the extracted certificate from your deployment manager server, to your plugin-cfg.xml directory.
- Open your httpd.conf file and check the location via the field WebSpherePluginConfig. In my case it's /opt/IBM/HTTPServer/Plugins/config/webserver1/plugin-cfg.xml
Open the plugin-key.kdb key database with iKeyMan.
- If it's not in the default location above, check the field by opening the plugin-cfg.xml file, and checking the field called keyring (i.e.: <Property Name="keyring" Value="/opt/IBM/HTTPServer/Plugins/config/webserver1/plugin-key.kdb"/>)
- The default password is WebAS
Select the dropdown "Personal Certificates" and select "Signer Certificates"
Click on "Export/Import..." and import the signer certificate that you exported from your deployment manager.
Restart the HTTPD and Administration servers
- /opt/IBM/HTTPServer/bin/apachectl stop
- /opt/IBM/HTTPServer/bin/adminctl stop
- /opt/IBM/HTTPServer/bin/adminctl start
- /opt/IBM/HTTPServer/bin/apachectl start

itim ACI refresh

Wed, 26 Oct 2011 15:16:15 GMT

When making changes to the TIM ACI's note that the default refresh time is 10 minutes. You can decrease the ACI refresh time by editing the enRole.properties file and changing the following parameter:

###########################################################
## AccessControlList refreshInterval - minutes to wait
## before cached ACIs are checked for changes and reloaded.
###########################################################
enrole.accesscontrollist.refreshInterval=0

Make sure it's set back to a reasonable value for production servers.

Installing TivSecUtil for TAMeb Combo Adapter

Thu, 15 Sep 2011 19:30:45 GMT

When installing on Windows 64bit you MUST install the 32bit version of GSKit 7.0.4.28. Installing the MSI for GSKit you get the error about Install engine being too old, and requiring an updated ISScript.MSI. You can get it from InstallShield's site (ISScript8.MSI). However, even after installing ISScript8.MSI the following error comes up when installing the Win 32bit GSKit:

---------------------------
Severe
---------------------------
usage: setup.exe <APP_NAME> [<ALT_TARGET_DIR>] [--logfile=[<FILENAME>]] [-SMS] -s -f1"<path>\setup.iss"

---------------------------
OK
---------------------------

Go into the directory where the setup.exe is and type "setup setup.iss". You are then able to install the Tivoli Security Utilities.

Readline support for DB2

Mon, 22 Aug 2011 20:39:29 GMT

So I got sick and tired of having to have to type in my DB2 commands over and over again. It would be nice to have readline support in db2. And thanks to rlwrap it now does. Readline wrapper basically emulates readline support for console based applications. As such:

touch ~/.db2_history
alias db2="rlwrap -i -f ~/.db2_history -H ~/.db2_history -s 1000 db2"

You can wrap any program like this. i.e.: xauth

touch ~/.xauth_history
alias db2="rlwrap -i -f ~/.xauth_history -H ~/.xauth_history -s 1000 xauth"

Facebook censoring books?!

Fri, 19 Aug 2011 10:09:33 GMT

Facebook won't let me add "The Buyout of America: How Private Equity Will Cause the Next Great Credit Crisis" as a book in my profile. This book is absolutely amazing. Private equity firms don't help ailing companies, they rape perfectly good ones by forcing them to take out loans they can't repay. Strip mine them of everything they have. All the while, leaving fund investors (aka pension funds, etc), employees, the gov't, and the company itself on the hook, while they walk away making millions. This book makes concrete examples and makes the lingo accesible to non-economistas like me. READ-THIS-BOOK!!

CityPaper websites has no security

Thu, 18 Aug 2011 15:19:27 GMT

I don't know who designed the City Paper.net's website, but it has little to no consideration on people's information security if you sign up for their site. To wit:
1) it stores and emails you your passwords in plain text (this is why I don't use Fring anymore either) This is what I sent to them.

I recently asked for a password reset. When I received the email I noticed that my password was returned to me in the clear. This is VERY BAD. This means that you are storing the password in your systems in plain text (or decryptable format) instead of a hashed format. This means that if anyone gets unauthorized (or even authorized) access to my data in your system, they will see my password.
...
This is a problem, because many people will use one password for many sites. If someone's password is compromised on your server, then they may have access to many other resources.
...

2)There is no way to change your password, change your contact information or name. Nothing.

What are they using an Excel Spreadsheet for their DB? Luckily I keep different passwords for different sites, which is a pain, but I do it for exactly this reason.

Criteria API ...

Wed, 10 Aug 2011 02:53:33 GMT

... because sometimes you have to make things more difficult for yourself than it is.

:-\

public void showSubquery() {
        _log.debug("Entering showSubquery()");
        Integer empNumberVar = new Integer(10019);
        
        /*
         * UM ... nuff said. :-\ Criteria API Sucks
         * 
         * select *
         * from salaries s
         * where s.emp_no = '10019'
         * and s.to_date IN 
         *  (select max(to_date) from salaries where emp_no = s.emp_no);
         * 
         */
        
        CriteriaBuilder cb = em.getCriteriaBuilder();
        
        CriteriaQuery<Salaries> mainQuery = cb.createQuery(Salaries.class);
        Root<Salaries> sal = mainQuery.from(Salaries.class);
        
        // Construct the subquery first, because we'll hve to use maxDateSubQuery
        Subquery<Long> maxDateSubQuery = mainQuery.subquery(Long.class);
        Root<Salaries> sqSal = maxDateSubQuery.from(Salaries.class);
        maxDateSubQuery.select(
                cb.max( sqSal.<Long>get("toDate") ) 
        );
        maxDateSubQuery.where(
                cb.equal(
                    sal.<Integer>get("employees").get("empNo"),
                    sqSal.<Integer>get("employees").get("empNo")
                )
        );
        
        // make sure you use parameters (sql injection protection)
        ParameterExpression pe = cb.parameter(Integer.class, "empNumberParam");
        mainQuery.select(sal);
        mainQuery.where( 
           cb.and(
              cb.equal(
                 sal.get("employees").get("empNo"),
                 pe ),
              // link your subquery
              cb.in(sal.get("toDate")).value(maxDateSubQuery)
           )
        );
        
        TypedQuery<Salaries> finalQuery = em.createQuery(mainQuery);
        finalQuery.setParameter("empNumberParam", empNumberVar);
        
        List<Salaries> results = finalQuery.getResultList();
        
        for (Salaries s : results) {
            _log.info(s.getSalary());
        }    
}

iPod classic transfer, crashing issues [SOLVED]

Sat, 23 Jul 2011 01:52:04 GMT

Two parts: 1) certain tracks which were once podcasts would not transfer to my ipod; 2) when clicking on podcast/playlist title with the play button, it would crash the ipod.

First, if you download music set (DJ set) podcasts, sometimes you would like to keep them. I created a Smart Playlist and added the necessary fields but they would not show up. Further, even when I added those tracks to the static playlist they would not transfer. I found out that I had to change the "Media Kind" from podcast to music, because I have the option set to only transfer unplayed podcasts to my ipod. You can set the Media Kind by "Open-Apple + i" clicking on the "Options" tab and setting the drop down to "Music".

Second, in order for my podcasts to play in chronological sort (oldest to newest) I would click on the title of the podcast or playlist with the Play button (not the middle one). I started getting the quirky behavior where when I do do that, it would flip out, go to the last track I played, tell me it was 0 minutes long, and if I pressed any other button the device would crash (immediately reboot). The problem was that I had it on Random. Apparently, it doesn't know what to do if it's on random, and you're telling it to play in chronological order... stupid computers. I switched it from Random to normal sort order, and everything worked as expected.

This was brought to you by the doesnt-apple-test-this-stuff-before-shipping department. Co-sponsored by the wheres-their-bugzilla department.

Sprint's 3G Bottleneck [Solved]

Wed, 13 Jul 2011 19:25:26 GMT

So I downloaded the latest android update and it seems like my 3G speeds went from 50k/sec to 8k/sec average. I finally figured out what it was. (Including listening to Public Radio apps etc.) They are pushing all HTTP and RTSP connections through a proxy server. So technically they aren't throttling the communications, but their proxy server just plain sucks. I found my old MSL code (I called them up for this way back when) and followed my own directions. Now I am back to normal speeds.
See: http://polywogsys.livejournal.com/284893.html
And: http://support.mog.com/kb/technical-support/fix-for-sprint-htc-evo-phones-after-3706511-software-update

For a brief synopsis:
Go to phone > Dialer > ##3282# > Edit Mode > Enter your "MSL" > You will see RTSP Proxy, HTTP PD Proxy > modify every list > set Address with "0.0.0.0", port with '0'.

Web Portlet Factory on Linux

Fri, 08 Jul 2011 21:55:39 GMT

So I decided to run the Web Portlet Factory on my linux desktop instead of a windows laptop. During installation, I selected an install directory, and selected the previously installed Eclipse 3.5 directory for the plugins. I then started to configure WPS7 server and ran into the following error:

com.ibm.websphere.management.exception.ConnectorException: ADMC0053E: 
The system cannot create a SOAP connector to connect to host 
portalServer.example.com at port 10025 with SOAP connector security 
enabled.

Doing some more digging, I noticed the following root error:


Caused by: [SOAPException: faultCode=SOAP-ENV:Client; msg=Error opening socket: 
javax.net.ssl.SSLHandshakeException: sun.security.validator.ValidatorException: 
PKIX path building failed: sun.security.provider.certpath.SunCertPathBuilderException: 
unable to find valid certification path to requested target; 
targetException=java.lang.IllegalArgumentException: Error opening socket: 
javax.net.ssl.SSLHandshakeException: sun.security.validator.ValidatorException: 
PKIX path building failed: sun.security.provider.certpath.SunCertPathBuilderException: 
unable to find valid certification path to requested target]

Apparently it was caused by my using Sun's JDK in eclipse instead of IBM's JDK. To fix this, I installed IBM's JDK and I edited my eclipse.ini file and added the following:

-vm
/opt/ibm/java/jre/bin/java

right above the "-vmargs" line.

Hope this helps someone it took me all afternoon to figure this out.

Netbeans Platform Refactoring Woes

Wed, 06 Jul 2011 21:32:20 GMT

So I am really not liking this new "Netbeans Platform" API. They have depreciated the Swing desktop application in favor of this new Netbeans Platform API. In any case, I created a new module and refactored the package name. When I did a clean/build all it failed with the following error: /opt/netbeans-7.0/harness/common.xml:252: Expecting localizing bundle: org/polywog/editor/api/lowercasefilter/Bundle.properties in: /home/exr003/Working/svn/programs/java/suite1/build/cluster/modules/org-polywog-editor-api-lowercasefilter.jar

Apparently the refactoring didn't pick up all of the references. In particular the reference to the Bundle.properties file. In the module which is broken, open "Important Files" and click on the "Module Manifest". In there you should see the reference to the old location of your Bundle.properties file. Change it manually in there to the new renamed location and it should build OK now.

Upgrading Fedora with /boot mdraid'ed

Fri, 01 Jul 2011 03:09:27 GMT

[Ok, I know the first thing out of your mouth is... why have a /boot in its own partition... we don't have to do that anymore! Well, it was a knee jerk reaction that is still biting me in the arse.]

In any case, this applies to any scenario where your bootable partition is mdraided and you want to upgrade Fedora (at least from F13 to F14). First off, let me just say that this is the LAST time that I will ever attempt to use "preupgrade-cli" (or preupgrade in general). It's cause me more headaches than it's worth.

If you have an mdraided drive that you boot from and want to upgrade Fedora just download the DVD and run through the normal upgrade procedures. it should succeed. Make sure you select to upgrade your boot loader too (just in case).

Problem 1) When you reboot it'll bounce you into the "grub>" prompt. I guessed (YMMV) which was the correct hd the bios was selecting, but you can guess too and/or try again - I'll explain in a moment.

I decided to hit this blog with the "grub" tag. Upon which I came across an entry where I wrote about "making sure your second drive is set up correctly in case your first drive fails and your only boot option is your second drive". I suppose I should have listened to myself. LOL. The article is polywogsys.livejournal.com/281987.html

Upon reaching the grub prompt, I just typed in the first two commands (I knew that my root or "/" was in /dev/sda3 and my "/boot" in /dev/sda2, while swap was in sda1). **NOTE** unplug all external - and unused internal - drives **NOTE**

Boot's root is where your grub/kernel files are (first hard drive, second partition) in other words "/boot" for ME. For you it might be just "/":

root (hd0,1)

Now install grub into the first harddrive (if that's the one the bios is using to boot from, if not, try hd1 instead)

setup (hd0)

I rebooted and was kicking and alive. I ran "yum -y update" and after a couple of more hours was alive and well with little downtime.

#ftw! :)

Slow DNS Response Time

Fri, 01 Jul 2011 02:48:04 GMT

So for the last couple of months, I have been having an issue where ssh'ing into my box would take over 6 seconds. Doing a "netstat -t" would take 7 seconds per external response. Figured it out... in my /etc/nsswitch.conf I had "hosts: mdns files dns" I remember putting that in because I wanted to take a bit of advantage of the protocol. Removing that, I now get the speed I need from my DNS responses. Very strange. But I ended up not using it so ... out it goes.

Upgrading from F13 to F14

Wed, 29 Jun 2011 20:24:33 GMT

Fedora 13 was just EOL. So I decided to upgrade my desktop at work to Fedora 14. My /boot was full so I couldn't do a preupgrade-cli. So I decided to boot the F14 DVD and do an offline upgrade. Note that I do have alternate repo's (for example rpmfusion and virtualbox) so I was hesitant. I upgraded OK, went for a reboot and everything seemed OK. I went to do a yum update and I got a PYURL error of "6" - Not very verbose. So finally I figured out that the problem was that networking was broken. I did an IFUP and ran dhclient on eth0 and everything after that seems to have worked OK. I have just booted into it and everything seems to still be working perfectly. Very smooth upgrade despite /boot hiccup.