As background, Fedora Linux versions are issued every six months or so, and no longer receive patches after version+2 is released. So I have been upgrading my system every two versions (every year) for the past several years. Particularly with respect to security vulnerabilities, it’s important to keep up to date with the latest patches.

So here’s a summary of my experience:

Distribution media boot problems

As I have done before, I downloaded the distribution DVD image via BitTorrent and burned a DVD+R. Last year the DVD booted smoothly, but this year it did not.  I tried several things to rule out DVD drive problems:  I booted last year’s Fedora 14 disk, and that worked fine, and I burned a Fedora 16 install disk using the drive on my MacBook Pro, and it failed to boot exactly the same as the one I burned directly on my Linux machine. My machine is somewhat older than my DVD drive, so it may not be a fully supported configuration, but that doesn’t explain why last year’s image booted OK. I unpacked the installation disk onto a partition on my network-attached storage, intending to use that for installation as I did two years ago.

So I downloaded and burned a network-install CD, which booted beautifully.  However, I got ahead of myself somehow and ended up doing the upgrade directly from the Internet.  It was perhaps a bit slower, but the upgrade went smoothly.

Graphics

As in the past, when the system did come up, the X Window System did not.  As suspected, this was due to my graphics card (NVIDIA GeForce 6200) not being configured properly on Fedora 16.  Some sort of a configuration problem has happened on each of the last two upgrades, so I was somewhat expecting this.  This time, I disabled the default nouveau driver and downloaded and enabled the NVIDIA proprietary driver.  Again no luck until I eventually recognized that the X11 configuration file, /etc/X11/xorg.conf, had the wrong name for the driver: “nv” instead of “nvidia”.  After also making sure that the ModulePath included the NVIDIA modules (/usr/lib/xorg/modules/extensions/nvidia), the display came up.

However, after logging in, Gnome 3 announced that my graphics driver wasn’t supported and it was operating in fallback mode.  I’m still not sure what I’m missing by being in fallback mode, but from reading some of the commentary on the Internet about Gnome 3, many people are happier with the fallback mode than the “real thing”. So I just forced fallback mode in System Info -> System Settings -> Displays, and I don’t get the annoying message any more.  Some day I may try to get out of fallback mode, but it’s not a priority right now, because of problems with…

Networking

With the exception of the Mac, my experience with user graphical environments has generally been that they operate at a “higher layer” than the network stuff underneath. They generally accept whatever they’re given by the network, and with the exception of some graphical tools to make configuration easier, don’t mess with the network configuration much.

That has apparently all changed with Gnome 3. Perhaps in an effort to act “Mac-like”, Gnome 3 attempts to reconfigure the network interfaces and creates new logical interfaces. Because my interfaces were set as “unmanaged”, Gnome (specifically the NetworkManager application) didn’t actually do this, but I had trouble getting applications like Firefox and Thunderbird to connect to the network. I reconfigured them to Managed, and suddenly I lost the static IP configuration I had. I had to manually reconfigure them to get back to “normal”.

However, I noticed that I was getting a LOT of syslog messages.  Every minute or so, syslog reported a configuration action including a number of errors such as:

NetworkManager[5700]: <error> [1325458737.135085] [nm-system.c:1061] nm_system_replace_default_ip6_route(): (eth2): failed to set IPv6 default route: -1

Despite the errors, IPv6 was working just fine, but I didn’t want the system generating all of these errors, so I disabled IPv6 and the errors stopped. I got busy with the search engines, and it appears that this is a problem that is shared with some other Unix variants, including Arch Linux and Debian. I haven’t seen any resolution to this; I have also submitted a question to the Fedora Forums regarding this problem.  It appears that every time my router sends an IPv6 route advertisement, this error cycle restarts.  It shouldn’t be doing anything unless the route advertisement has changed.

Following the disabling of IPv6, I got a lot of errors from named (the BIND name server) not being able to reach IPv6 addresses in the course of resolving addresses. I added the -4 option in /etc/sysconfig/named to cause it to only use IPv4 for the time being.

Bottom line: IPv6, for me at least, is basically unusable in Fedora 16.  This is a serious deficiency.

Services

Fedora 16 has changed the way that system services (daemons, etc) are managed.  While in the past this was done directly with the service and chkconfig commands, this is now done with the systemctl command.  Fortunately, service commands are still accepted and translated into systemctl commands for those of us for whom habits die slowly.

One of the first things I noticed on upgrading to Fedora 16 is that dovecot, the IMAP server I use, no longer worked. It turns out that the service was disabled for some unknown reason. Re-enabling and starting the service solved the problem

But after a day or so I noticed that I was no longer receiving logwatch emails. It turns out that the crond process had similarly been disabled and was not running. While dovecot is something that not everyone uses, I am really puzzled about why cron would have been disabled.

Summary

This year’s Fedora upgrade involved about the same level of problems as previous years, with some chronic problems (boot, graphics) and some unique problems (networking, services). I am quite unhappy at having to disable IPv6 to keep the system from reconfiguring the network every minute or so; I hope this problem is fixed soon. Otherwise, this seems to have been an OK upgrade, but I’m glad this only happens once a year.

When you’ve been at a company for such a long time, it’s easy to take for granted all the things one’s company does right and focus on the things it does wrong. At its core, Cisco sets a very high standard for honest and legal business conduct. Cisco encourages its employees to give back to the community, both through incentives like matching grants and by giving employees space to allow them to contribute their time and talent. I have been proud to call myself a Cisco employee.

I liken Cisco to an aircraft carrier: it’s very big and powerful, and shows up on all the radar screens. It has a lot of firepower at its disposal; I wouldn’t want to be in its crosshairs. But it’s also difficult to stop, turn or otherwise maneuver. The possibility of a member of the crew influencing its direction of the ship is virtually nil.

A zodiac boat, on the other hand, is all about speed and maneuverability. It’s a lot easier to point at something on the horizon and suggest to the Captain that we should go in that direction. Zodiac boats often operate “under the radar”, heading in a new direction amid the clutter of other vessels.

On an aircraft carrier, it is actually possible to go in a slightly different direction from the ship for a while.  You do this by walking from one side of the deck to the other. But you’re well advised to stop before you get to the edge.  I liken this to my recent experience at Cisco:  my career direction was leading me off in a slightly different direction.  Cisco was great in letting me explore that area (identity management), but there was only so far that I could go with it there.

Fortunately, I found a nearby zodiac boat, OneID, moving in the direction I want to travel.  The captain (CEO) is an successful serial entrepreneur, Steve Kirsch, that was a classmate at MIT. I’m looking forward to the trip, even though I realize that I’ll no longer have a lot of the comforts of being on an aircraft carrier.  The trip might be rougher, but I’ll be going in the direction I want. I’m looking forward to the journey.

Let me start by saying that I strongly support the need for anonymity and pseudonymity on the Internet. There are many users that would not feel the freedom to express their opinions freely if their identity is associated with their expression. For many of them, there is good reason for this as in some countries people are jailed or otherwise persecuted for opinions that differ from, for example, that of their government.  The United States has a long tradition of anonymous political discourse, and there are other situations, such as whistle-blower programs and crime-tip hotlines, where anonymity and pseudonymity provide an easily demonstrable public good.

The question here is whether Google+ should be able to require its users to use their “real” names on the service, not whether attribution is required on the Internet as a whole.  One could characterize Google+ as more like a place of business that is exercising its prerogative to set the rules for doing business there. If you don’t want to abide by the rules, go elsewhere.

My possibly unpopular opinion is that services on the Internet should decide what information they require of their users, and users should decide whether or not they are interested on that basis. Google+’s profile names policy has said that they require a user’s real name, and if there is a segment of the population that isn’t willing to abide by that, they should find another one that meets their needs.  They are free to make their opinions known to Google and others, but it’s really Google’s call.

I’m not sure whether this stance on Google’s part is a good idea or not: they are effectively saying that people who want to use pseudonyms on their social networking service are not welcome there. Social networking services grow in value nonlinearly based on the number of users they attract, and Google+ is basically saying “go away” to some subset of them. But Google+ is a work in progress: they have said that commercial pages will be added in the future, and one of the motivations might be to discourage commercial pages for characters like “Joe Camel” or “Ronald McDonald” from being created.

On one of the mailing lists to which I subscribe, Mary Hodder expressed the opinion that Google’s (and Facebook’s) market power make them more like the Mall of America, where courts have ruled that they must be treated as public places.  I’m not nearly well enough versed in the law (nor in the specific rulings) to know whether that applies here, but to me it seems like a bit of a stretch for a service as new as Google+ to be considered a public place so soon.

However, Google+ has been far from perfect in the implementation of their profile name policy:

• Some users with unusual (in many cases non-Western) names have been suspended
• Undoubtedly many people using pseudonyms that look like typical Western names are getting away with it
• Appeal procedures have been uneven.  Many have been asked to prove who they are by providing a scan of a Government ID document
• The definition of allowable variations of legal name is fuzzy.  The Google+ policy says that “if your full legal name is Charles Jones Jr. but you normally use Chuck Jones or Junior Jones, either of those would be acceptable” but it’s a lot less clear if you normally use a different last name.

I wouldn’t be surprised if Google was caught unprepared for some of these issues, or perhaps by the volume of complaints.  The fact that people are complaining bitterly about the use of pseudonyms rather than simply leaving is a great compliment to the Google+ service; in some respects it is reminiscent of the New Coke debacle of the ’80s.

This is a very timely debate.  The National Strategy for Trusted Identities in Cyberspace (NSTIC) has as one of its premises a very market-driven idea:  that the relying parties (internet services you use) can specify what information they need from users, and users get to decide what “attributes” (personal information) they release to them. The availability of a trustable source for a user’s legal name would solve some of the issues above (although not the allowable variations problem).  But would users have sufficient market power to decline to provide that information?  It is the tussle over that market power that is at the heart of the “nymwars” issue.

Creative Commons-licensed image above from user clydesan on Flickr

On this last day of our trip, we again have an afternoon flight so we had a little time to walk around downtown Anchorage a bit more. As on Day 2, our first choice for breakfast had a long wait, but we found a quite reasonable alternative. We had seen a sign for All Alaska Tours, and we thought we would stop by and tell them what a great trip we had. When we got there, they had moved a few blocks away so we treated it as an opportunity to see some more of the city. We were surprised about the stores selling fur clothing, which is definitely not done around home.

A sign in Anchorage's Earthquake Park. We like their sense of humor.

We said “hi” and thanked Jessica, our travel arranger, who was probably startled by our unannounced visit but handled it well. We returned to the hotel to finish our packing and to try to figure out to do with the leftovers from our picnic lunches. We loaded the car, and having a little time to spare, stopped by Anchorage’s Earthquake Park. Here we learned a bit about what happened in Anchorage as a result of the 1964 earthquake, although we kept our visit short so as not to provide too much food for the mosquitoes.

Our flight was delayed about an hour and a half due to the late arrival of the incoming plane. Ted Stevens International Airport in Anchorage has a modern and attractive terminal, featuring free (and reasonably fast) WiFi and a few good shops, so the time passed quickly.

The flight went uneventfully.  We collected our bags, went outside and waited for the parking shuttle. It was dark, for the first time in two weeks.  We are home.

We awoke today to weather that is apparently more typical for Seward: low clouds.  It made it easy to forget all the mountains that are nearby. We packed, checked out of the hotel, and headed to the nearby Subway sandwich shop for breakfast that was quite passable.

Junior Ranger Ceremony

After breakfast, we returned to the Kenai Fjords National Park visitor center so that Celeste could turn in her booklet for her Junior Ranger badge.  Junior Ranger is a program the national parks have for kids 4-12 that encourages them to learn and explore, and to ask good questions. We had picked up a Junior Ranger booklet on the our first visit to the Exit Glacier visitor center. While Celeste knew the answers to many of the questions already, there was also an assignment that required her to interview a Park Ranger, and encouraging that kind of contact between kids and the rangers is always a good thing.  She took an oath to protect the park (they always make it a good ceremony) and was given a small badge.

We then drove to the town of Whittier, a former military town.  While not far off the main highway, Whittier is on the coast on the other side of a mountain, and is connected only through a one-way tunnel that is shared by vehicles and trains.  There is a toll ($12 to go, free to return) and one waits in a queue not unlike that waiting to board a ferry until the tunnel is open in the direction you wish to travel.

Train entering the shared tunnel at Whittier

When we got to Whittier, there was heavy mist and it was fairly wet. We walked through its small waterfront with a few shops and restaurants and then through a tunnel to downtown.  Downtown Whittier is not very attractive, with many derelict boats and a single high-rise condo building that nearly everyone lives in. The high-rise is a cold-war era military barracks built of heavy concrete to withstand both the harsh conditions and possible earthquakes and attack.  We walked over to an adjacent building that housed the Whittier Museum and saw pictures of the aftermath of the great 1964 earthquake. Whittier was devastated by both the earthquake and by a tsunami that followed soon thereafter. Very little was left.

Returning from Whittier, we had our picnic lunch in Girdwood, a bit further toward Anchorage. Girdwood was preparing (girding?) for a town festival for the July 4 weekend, so we found a convenient parking place and had our picnic out of the back of the minivan.

Returning to Anchorage, we checked in again at the Marriott we had stayed at on Day 1. We went off to the Snow Goose where we had eaten that first evening, and all remarked on how quickly the two weeks had gone.

Despite the weather forecast, today started off mostly sunny. We opted for breakfast at the hotel, which was pretty good but (as usual) marginally too much food.

Exit Glacier Nature Center and Exit Glacier

We went back to Kenai Fjords National Park for the 10 am ranger-led hike to the edge of Exit Glacier. As usual, the ranger (a young woman) was very informative and was able to answer all of our questions well. At the end of the hike, we found out that this is her first season at this park; like many of the seasonal rangers, she moves from park to park as required, and has previously been at Pinnacles and at Lava Beds National Monument. It makes her knowledge of the park all that much more impressive.

Unlike the Root Glacier that we visited earlier in the trip, there was a deep crevasse between us and the glacier itself, so it was not possible to walk on it this time. Nevertheless, the blue glacial ice color is worth the trip by itself. We also learned a lot more about how the forest establishes itself on ground freed by the recession of the glacier.

Today’s picnic lunch was on the front balcony at our hotel room. In addition to the sizable rooms, they have tables and chairs that were suitable for us to set up our spread of food and sit very comfortably.

After lunch, we ventured back into downtown Seward. Seward has two parts: the southern “old town” and a somewhat more touristy district further north near the marina. We started with the old town, where Shirley, Kenna, and Celeste had a great time at a quilt shop, Jim found an electronics store, I found a great cafe (with Intelligentsia coffee!), and we all loved the bookstore.

Seward Harbor

We then moved to the northern district, visiting the Kenai Fjords visitor center (excellent films), then checking out some souvenir shops and surveying possibilities for dinner.  When it came time for dinner, our first choice, a Mexican cantina, was too crowded, so we opted for an adjacent restaurant that worked out fine.

The weather forecast for the rest of our trip all seems to be the same: mostly cloudy with occasional showers. We have done pretty well with weather so far, and hope our luck holds out. This morning, as has frequently been the case, the weather seems somewhat brighter and nicer than the forecast would imply.

Main Street, Hope, Alaska

Breakfast was delivered to our cabin, and consisted of juice, pastries, and yogurt; all excellent. Then we took off for Seward, stopping off to replenish our picnic supplies. After passing through Anchorage, we enjoyed the very scenic drive on the Seward Highway along Turnagain Arm. The weather varied from partial sun to rain along the way.

We decided to make a side trip to Hope, Alaska, in part because Hope is Celeste’s middle name. Hope is a charming old gold mining town that was devastated (considerably reduced in size) by the great earthquake of 1964. We found a great place for a picnic lunch, and then walked around the town, visiting the museum and a couple of gift shops. The museum tour was particularly informative, as we were led around the buildings by “Billy”, who we would estimate to be at least in his 80s, who kept referring to the “old timers” he learned so much from.

Hope Museum exhibits

Our hotel, the Seward Windsong Lodge, is just north of town on the road to the Exit Glacier area of Kenai Fjords National Park. Our rooms were enormous, a welcome change from many places where we have had to step over our suitcases to get around.  After checking in, we made a trip out to the park to scout out hiking opportunities for tomorrow.  We decided on a 10 am hike to the edge of the glacier. We then headed downtown for dinner and a little souvenir shopping.

The day started, as many have, cloudy with a threat of rain. We wanted to revisit the store at the Denali Visitor Center, so we opted to have breakfast there as well. The choices were limited, and delivery time for those who ordered a hot breakfast was very slow.

This was one of the longer drives of the trip. This part of the Parks Highway was very scenic,  but we were plagued by slow-moving mega-RVs that were sometimes difficult to pass and most of which (particularly two with Arkansas plates) did not pull over when a line formed behind them. We stopped from time to time to look at things, even though it meant that we had to pass some of the same RVs a second time.

One of these stops was at a rather bizarre cross between a hardware store and a flea market called Wal-Mike’s in Trapper Creek. They sold Alaska license plates with just about anything you wanted on them.  There were also a lot of references to Sarah Palin, things like old campaign signs that were for sale, but we decided not to ask about her.  Picnic lunch was at a visitor center at the turnoff for the Talkeetna spur road. Although the visitor center was closed this time, the picnic tables were fine until it started to rain.  We had off-and-on rain much of the day.

Matanuska Experiment Farm

After a quick drive through Wasilla (mostly harmless) we stopped at the University of Alaska’s Matanuska Experiment Farm just southeast of town. We explained that we were tourists and got a wonderful introduction and walking tour of the facility, led by the superintendent.  The work they’re doing includes efforts to try to understand losses in the moose population, use of local grasses for hay to feed livestock, tests of various potato yields in local conditions, and extraction of synthetic oil from wood pulp. Unfortunately this facility and others like it are losing a lot of their funding because agricultural research doesn’t have the popularity of many other Federal programs.

Our lodging for the night is the Knik River Lodge, south of Palmer and about an hour’s drive from Anchorage. Despite its relative proximity to Anchorage, you would think we are out in the middle of nowhere: it’s very quiet and secluded. The rooms are individual cabins on a hillside overlooking the Knik River Valley. Dinner was very delicious and served in a large circular tent (a yurt). We were happy to reconnect with a group from the Netherlands that we had encountered in Kennecott earlier in the trip.

Knik River Lodge cabins

We had booked an all-day (actually 14-hour) bus tour down most of the main road of Denali National Park. To minimize congestion and for preservation of the park, private vehicles are for the most part not permitted on the park road. Instead, there is a bus system for transporting visitors. In addition to the scheduled buses, there are tours operated by concessions under contract to the National Park Service. Our tour took us basically the entire length of Park Road, a distance of 92 miles each way.

Our bus driver/tour guide, Kevin, did an excellent job of describing the history of the park and of spotting and stopping the bus for many wildlife sightings. Among the animals we saw were moose, caribou, Dall sheep, grizzly bears, and golden eagles. But one of the most prized sightings was not of a living thing: the sky cleared just enough, at the right time, for us to see Mt. McKinley! We have talked with quite a few Alaska residents who haven’t had that opportunity yet.

At the midpoint of our trip, we stopped by Kantishna Roadhouse for lunch and some activities. We chose the dog sledding demonstration, where I learned that the predominant sled dogs aren’t huskies any more, but rather mixed breeds chosen for their temperament and physical characteristics.  Irish setters, for example, lend their excitement to the team very favorably.

Click to view slideshow.Returning to the start brought more animal sightings, particularly grizzly bears. After a light dinner at the hotel, we turned in early.  14 hours on dirt roads takes a lot out of a person.

Park on the Chena River in Fairbanks

Today’s drive is relatively short, so we took some time exploring downtown Fairbanks and attending to some grocery shopping and errands.  While much of downtown Fairbanks isn’t all that attractive, the riverfront has been made into a very nice park with exhibits from Alaska’s 25th anniversary of statehood in 1984.

Proceeding down the Parks Highway toward Denali, we passed through the town of Nenana, which seemed like a good place for the day’s picnic lunch. The visitor center there was a real treat. We stopped because of the public picnic tables, but after a while the volunteer there came out to offer us coffee. When we went to talk with her and learn about the area, we found out about the Nenana Ice Classic, an annual contest to predict the date of the annual ice breakup on the Nenana River. She also had a wealth of information about the Denali area and onward.

Nenana Ice Classic "Tripod" for detecting ice breakup

Arrival at Denali greeted us like so many national parks, with a hyper-commercial business district just outside the entrance. Hotels, restaurants, souvenir shops, tour operators, and the like were well represented. The district definitely earns its nickname, Glitter Gulch. We found the way to our hotel, the Grande Denali Lodge, up a steep dirt road with switchbacks. The hotel seems very nice, again the type frequented by tour buses, and our rooms have a wonderful view (of the east end of the park and Glitter Gulch).

It was still mid-afternoon so we went to the park Visitor Center, viewed exhibits, and took a short hike. Yelp helped us find a reasonable place for dinner and then we retired.  Tomorrow we need to be ready to be picked up at 6:00 am for our tour of Denali Park, so it’s going to be an early morning.