Vulcanair 1.0 Training Airplane

Vulcanair 1.0

I wasn’t aware that Vulcanair bought Partenavia out of bankruptcy in 1998, and has built an EASA-certified Cessna 172 clone, the Vulcanair 1.0, for almost half the price.

The 1.0 is a corrosion-proofed metal G500-based glass-cockpit 4-seater with an IO-360 engine. It has roughly the same performance as the old 172 XP or new Cessna SP, IFR-equipped for $259,000.

It has a constant-speed propeller but fixed landing gear, so doesn’t qualify as a complex aircraft. It would feel sporty when flown solo, climbing at over 1,000’/minute.

The 1.0 was announced in 2014, was exhibited this year at Oshkosh (AirVenture) and is targeting Fall 2017 for FAA certification. I didn’t see any listings on or, so there aren’t many flying yet. Vulcanair 1.0 Vulcanair’s Skyhawk Competitor Vulcanair V1.0 Trainer Aiming at U.S. Training Market
Vulcanair Brazil Prepares for P68 Assembly The U.S. has a staggering pilot shortage
One Aviation Unveils Testbed with Eclipse 700 Wing
What’s The Best Flight School Trainer?

Posted in Tech | Leave a comment

PSA: Workarounds for CentOS 7.3 Problems with SE Linux 2.5

PSA: If you update to CentOS 7.3 and see odd console or log errors like this resulting in a hung boot:

Failed to start Import network configuration from initramfs


work still pending
FAILED Failed to start Login Service.
See 'systemctl status systemd-logind.service' for details.
FAILED Failed to start Authorization Manager.
See 'systemctl status polkit.service' for details.
DEPEND Dependency failed for Dynamic System Tuning Daemon.

then you have a problem with CentOS 7.3 and SE Linux 2.5.

The workarounds are surprisingly simple:

  1. don’t upgrade anything until CentOS 7.4 is released (verified by me on Sept. 17 on a Dell 1950) or
  2. add selinux=0 to the kernel boot line in the boot menu and/or grub. Startup fails with multiple errors

Keywords: Virtualbox, linux, selinux, enforcing.

Posted in Linux, Tech | Leave a comment

Air Canada 759 Near-Miss at SFO on July 7, 2017

On July 7, 2017 an Air Canada jet nearly landed on a busy taxiway, Charlie, and had to do a go-around at SFO. It missed another airliner on the taxiway by less than 100′, and other airliners behind the first.

There’s three problems with that:

  1. the pilot was cleared to land, not cleared for a low approach over taxiways
  2. Low approaches over taxiways can only be approved at 500′ above ground. See “3−10−10. ALTITUDE RESTRICTED LOW APPROACH”: “A low approach with an altitude restriction of not less than 500 feet above the airport may be authorized except over an aircraft in takeoff position or a departure aircraft. Do not clear aircraft for restricted altitude low approaches over personnel unless airport authorities have advised these personnel that the approaches will be conducted. Advise the approaching aircraft of the location of applicable ground traffic, personnel, or equipment.”
  3. jet wash at TOGA power can be destructive, especially against the flight controls of the airplanes on the ground.

The go-around is a normal procedure, and in fact your only 2 options are landing or go-around, so that wasn’t a problem, or even unusual, as some reporters wrote.

At this time we’re fairly sure the pilot and copilot couldn’t identify the correct runway, but the question is why not?

Some possibilities:

Category Risk Factor
Aeromedical fatigue, as the flight was late at night, especially considering the departure was from the East Coast
Aeromedical QUIET BRIDGE VISUAL RWY 28L/R approach was flown using FMS, possibly abstracting the flight path in a confusing manner than a relatively simple ILS
Aeromedical approach was visual at night, much riskier than VFR Day. I believe European airlines don’t allow this.
Facilities Runway 28L was closed for construction and the normal lights turned off, changing the expected landing sight picture
Facilities taxiway lighting might be LED, thus appearing white instead of blue (taxiway edge) or green (taxiway center) at a distance
Navigation ILS is often inoperative at SFO (likely the actual cause of 2013 Asiana Flight 214 accident)
Controller tower/ATC is either not adequately engaged, or needs more radar or cameras. It’s great that the UA pilot radioed, “Where’s this guy going? He’s on the taxiway.” but it would have been even better had Tower noticed that first.

The reason for the current excellent airline safety record is having scheduled flights from Point A to Point B, which are known in advance (ie. scheduled.) However, SFO is an unsafe airport that has several surprises for pilots:

  1. old US airports like SFO were built with parallel runways too close together for safe parallel IFR operations
  2. SFO is between 2 mountain ranges (San Bruno and Diablo)
  3. SFO is next to a “black hole” (SF Bay is not visible at night)
  4. increased navigation risk with continual ILS problems
  5. now add construction and lighting changes.

SFO is more of an obstacle course than a safe airport at this point. The first 3 problems are inherent in SFO’s location, but the last 2 are airport-management induced.

The Air Canada cockpit voice recorder was not stopped, and eventually recorded over this time period as the NTSB did not learn of the incident for 2 days. Ironically, the most valuable information is probably the routine cockpit conversation about the approach setup 3 miles out, (the why?), not the last-second go-around scramble.

The NTSB will probably try to estimate the distance. It should be interesting how they do this calculation because of how small the times and distances are. Normally aircraft tracks are not calculated to the foot!

Also of PR interest is that the aviation community waited for more data on the loss of separation distances before commenting, while the press and flying public were quick to outrage.

One forum poster called this “closing ranks,” which is not a fair criticism since airplanes often approach to the “wrong runway” (daily examples are circle-to-land procedures and Tower change of parallel runway instructions) and overfly other planes in the airport area.

In fact, Tower prefers transiting VFR airplanes overfly the center of the airport because then they know precisely where you are, versus interfering with landing or departing paths.

Disclaimer: I hold a commercial airplane licence and have flown Bay Tours through the SFO airspace, but I am not an airline pilot. Investigators probing how Air Canada pilots’ mistakes led to near-disaster at San Francisco airport Disaster Averted! Or Not? Near miss with 5 airliners waiting for T/O on taxiway “C” in SFO! Exclusive: SFO near miss might have triggered ‘greatest aviation disaster in history’, HN
NTSB Issues Investigative Update on San Francisco Airport Near Miss
Air Canada disappeared off SFO air traffic radar equipment for 12 seconds before near-disaster, NTSB says Air Canada Flight Misses By Four Feet?
SFO Landing Incident Prompts Focus on Pilot Monitoring, CRM

Posted in Tech | Leave a comment

Solr Meetup at Cloudera in Palo Alto

Cloudera hosted another Solr Meetup at their office in Palo Alto. About 30 people attended.

Two software engineers from Cloudera did presentations tonite:

1) Michael Sun talked about his nitely Solr microbenchmark and cluster benchmarks.

Solr Nightly Benchmarks (SOLR-10317)

2) Mano Kovacs from Cloudera talked about Solr LIR trouble-shooting techniques

Nice t-shirt: “Data is the New Bacon.” :)

Leader incorrectly publishes state for replica when it puts replica into LIR (SOLR-9555)

He also talked about limitations and issues of of autoAddReplicas.

Recommends PlantUML for documentation.

Audience Questions

– piercing – checking 100 document versions may not be enough if there’s 1000 total versions

After party at Antonio’s Nuthouse on California Ave.

Thanks to Cloudera for hosting this event and the Mediterranean food!

Cloudera 1001 Page Mill Road Palo Alto, CA 94304

Apache Solr Memory Tuning for Production Algolia raises $53 million for its search engine API

Posted in Cloud, Storage, Tech | Leave a comment

Distributed Systems Laws Applied to Distributed Databases

Perl LogoAvery’s Law of Distributed Systems Reliability: “Distributed systems are more reliable when you can get a service from one node OR another. They get less reliable when a service depends on one node AND another. And the numbers combine multiplicatively, so the more nodes you have, the faster it drops off.”

Lamport’s Observation: “A distributed system is one in which the failure of a computer you didn’t even know existed can render your own computer unusable.”

Both sound simple enough, even obvious for systems serving HTTP(S) or caching.

But apply those to clustered and distributed databases, and you can do powerful analysis on expected availability.

Topo Nodes Database Configuration Reads during single node failure Writes without a failure Writes during single node failure
WAN 2 MySQL Master-Slave Async Replication OR Single Master NA
WAN 2 MySQL Master-Master (standby) Async Replication OR OR (Single Active) OR (Single Active)
LAN 2 MySQL Master-Slave Semi-sync Replication OR Single Master NA
LAN 2/4 MySQL CGE NDB Cluster Sync Replication OR OR NA
LAN 3 (Min.) MySQL Galera Sync Replication OR OR Requires quorum for OR
WAN 3 (Min.) Cassandra Appropriate RF and CL OR OR OR Turing Award to Leslie Lamport

Posted in Cassandra, Cloud, Storage, Tech | Leave a comment