Reading the fine print

Upon contemplation, blaming a time zone issue for the mail server problem just doesn’t seem right. Email works just fine across all the various time zones; why should it fail over something so simple?

When I “solved” the problem it was by updating the time zone, and bringing the server into proper sync with the client. But why would this be needed? As I made the change, I noticed something else unusual – there was a persistent connection to my mail server’s IMAP port, from the Amazon cloud. That struck me as odd, and potentially undesirable. So I closed the connection, and waited… and it reestablished itself within seconds. I put this down to being the tech-support diagnostic from the email client vendor, and then went about the business of kicking them off the server (write email to them, remove the extra account created for them, changing master passwords). As always, when I’ve opened a diagnostics port for outsiders, when it’s finished I go back and reset any master passwords which might have been [inadvertently] exposed – in this case, the password to my primary email – access to which had been the impetus for the whole exercise.

Changing the password on that email requires updating not just the mobile client on the phone, but also the mobile clients on the notebook computers, and the POP3 client on the administrative computer.  It’s an easy routine; pull up the client, stop the automatic message pull (which will fail as the password has changed), then make the change, do another message pull to validate, and we’re done.

Not so, this time. Instead, upon updating the password in the POP3 client, I get a “server closed connection” message. What? Did I type the password wrong? Let’s go try that again… get the same error… but that’s the wrong error for a password change. So it’s off to the mail server console to see what’s happened now – and the primary email account is now on ‘auto-ban’ status, meaning too many attempts were made with the wrong password in too short a time frame. I remove the account from auto-ban, do the POP3 pull from the client (works fine), then check the phone app.

It’s complaining it can’t connect. And that’s when it hits me – that persistent cloud connection was from the phone app. But why would it care? I’ve told the phone app to poll the server every 15 minutes… let’s fire up wireshark and do some packet analysis, see what’s happening. And sure enough, it’s the phone app, probing the mail server every 60 seconds for new email. Say what? And then I notice the section on the app menu, about privacy, and look in there. “Erase your data from our servers” – ahem. What data are you storing? Selecting Yes to “Erase data” tells me I may no longer use this app, and when I persist, it deletes the account from the app – resetting it back to the beginning.

Let’s go examine that app again… I should have been a bit more aware (read the fine print!) as to what was going on; how did a no-ads-install-for-free app have a dedicated tech-support team, without a visible source of revenue? Well, if you’re not paying, you’re the product, not the customer. In this case, the company providing the app is indeed reading the mail and making marketing trend analysis data available – to third-party [paying] clients. That’s how the app makes money.

I’d complain about deceptive advertising but… GMail is essentially the same. So Edison Email comes off the phone, and it’s time to find a better client.

Advertisements

Time and the ancient email server

For the past few days I’ve been chasing a rather pernicious problem – email clients weren’t interacting properly with the server. Since I haven’t changed the server software in, oh, five or six years, my first thought was – there’s something crazy going on with the clients.

As of this writing (November 2018) I run my own primary email server. It’s been a stable platform for years, requiring only very intermittent maintenance. However, it’s an older product, running on a non-supported platform. I run in a blended mode; the server is nominally IMAP4 but when I’m in the office I’m retrieving messages using the older POP3 protocol – which then removes the messages off the server. I’m used to this routine (23 years and counting) and it limits message exposure to only a few hours under most circumstances.

The symptom was – the phone email client showed unread messages on the server, but gave me no way to access them. Over the period of a few days I narrowed it down to a specific sequence – leave a few messages on the server, off to bed, wake up in morning, check mail and find the unread and inaccessible new mail… which then caused me to do a POP3 run and retrieve stuff. But this wouldn’t do for the long run. I experienced the same phenomenon on Edison Email (Android) and on GMail (browser and Android client).

The Edison Software app has excellent tech support – and during the various steps they went through (including establishing a test account on my server) – I finally recognized the problem. My server had already executed the reset to standard time – two weeks early. Once I reset the server back to daylight time, things are back to working normally. Then it was time to track down the timezone file and fix that, so this won’t be a problem (at least for the next several months).

 

Sunday morning on the rails

The good people of the Black River Railroad Historic Trust decided to run a steam-powered mixed-freight train. Naturally I had to be there, to get some photos.

grey coal smoke rises from the woods
Smoke rises from the woods as Engine #60 prepares for its time on camera.
steam engine on the move
First photo run of the morning; #60 pours on the power.
train crew watching
A careful crew is a safe crew; watching the track ahead.
engineer's hand on throttle
Hand on the throttle.
group of photographers trackside
Photographers and the conductor await the run at Bowne.
steam engine making turn
Cue the music as she comes around the bend.
Steam train gets closer
More power.
Steam train with observers in foreground
On the second run-by.
train crew talks with conductor
Crew receives new orders
Caboose with crew members
Starting the shove back home.

Several times during the trip, I was asked if I wanted to stand right up against the right-of-way (ROW in rail terms); that’s not necessary for my interests. It was a great way to spend a few hours on a sunny Sunday in September.

 

Off the schedule

Come January 2019, I’ll have all the nights of the week free – as of today I’m  off the schedule for Spring 2019. It’s the end of the line for teaching… 21 years in higher education is enough.

Soon I’ll start the process of official retirement for the NJ PERS (Public Employee Retirement System), so that I can collect a pension check… while NJ still has money to pay.

I’m going out on the horse I came in on. I was originally hired to teach the ‘hardware’ course – and that is the course I’m teaching now. Symmetry.

 

Leaving Facebook… the process.

I don’t intend (at least in the next year or so) to actually leave Facebook. It’s a nice way of keeping in touch with some people, and a good experimental platform for understanding the ways and wiles of tracking systems.

But there’s been a lot of chatter the last few days about how to leave Facebook, and the [obvious, if you think about it] algorithms used by Facebook for victim customer advertising target retention. Specifically, Facebook starts hitting the notifications hot and heavy, and eventually starts ignoring its own settings in the quest to keep you entertained on the hook.

The process you should follow (at least as of this writing)… 1) Start removing your content. Don’t do it all at once, but my approach is to slowly remove all photo “albums” with the exception of the auto-generated foursome (uploads, timeline, profile, cover).

2) Create an alternate identity email somewhere. Change your Facebook email over to this new account. Don’t use this email for any other purpose.

3) Remove ALL references to your primary email, web, phone, etc. from Facebook. Don’t use any identifier other than the alternate email address.

4) After a month or so, disconnect Facebook from your phone (if you ever gave it the number to start with)… watch as Facebook starts to fill your auxiliary email with notices.

5) On your fated day, sign in to Facebook, do the removal thing – it may take a number of attempts complete with various captchas and pleadings and confirmations and so on.

6) You may now walk away (figuratively speaking) from that alternate email, safe in the knowledge that Facebook can’t bug you any further.

 

 

Winding down…

I’ve started winding things down. This blog will continue as an intermittent feature, but its days of being hosted on wordpress.com are numbered. GDPR and a poor choice of username are the main culprits for this change.

It’s time to move on… 34 years in the same gig is long enough, time to find that greener pasture elsewhere. The future will emphasize photography, travel, and writing; software development and education will take a back seat.

The main sites will be moved to virtual private servers out in the cloud – the days of hosting internally will come to a close sometime late this year. A few things will not make the transition.

 

Reading instructions…

There’s just two more sessions left this term for my “Internet & Web Architecture” class (tonight, and next Friday night).

The class has a weird name. I preferred calling it “Introduction to Systems Administration” or perhaps “Basics of Internet Infrastructure” but when you’re the adjunct (part-timer) and the PhD wants another title… you go with what the PhD wants.

Either way, the basis of the class is to teach the bits and pieces most classes ignore – setting up servers, virtualization chores, DNS, registering domains… and introduces students to a real-world issue: keeping track of credentials.

It’s this last bit which causes the problems… that, and that no one seems to want to read instructions.

Well, guess what. I’d rather not read them, either. Except I know what happens when you don’t read the instructions – it blows up in your face, and then you have to go and start all over again, and this time, read the instructions.

So reading instructions first turns out to save time.

But just try getting students to understand that…