For a long time I’ve been a supporter of Virgin Media (from a broadband perspective, though their triple play TV/Phone/Broadband offering has seemed decent too). I know they have a bad reputation amongst some people, but I’ve always found their engineers to be capable, their service in general reliable, and they can manage much faster speeds than any UK ADSL/VDSL service at cheaper prices. I’ve used their services everywhere I’ve lived that they were available, starting back in 2001 when I lived in Norwich. The customer support experience with my most recent move has been so bad that I am no longer of the opinion it is a good idea to use their service.

Part of me wonders if the customer support has got worse recently, or if I’ve just been lucky. We had a problem about 6 months ago which was clearly a loss of signal on the line (the modem failed to see anything and I could clearly pinpoint when this had happened as I have collectd monitoring things). Support were insistent they could do a reset and fix things, then said my problem was the modem and I needed a new one (I was on an original v1 hub and the v3 was the current model). I was extremely dubious but they insisted. It didn’t help, and we ended up with an engineer visit - who immediately was able to say they’d been disconnecting noisy lines that should have been unused at the time my signal went down, and then proceeded to confirm my line had been unhooked at the cabinet and then when it was obvious the line was noisy and would have caused problems if hooked back up patched me into the adjacent connection next door. Great service from the engineer, but support should have been aware of work in the area and been able to figure out that might have been a problem rather than me having a 4-day outage and numerous phone calls when the “resets” didn’t fix things.

Anyway. I moved house recently, and got keys before moving out of the old place, so decided to be organised and get broadband setup before moving in - there was no existing BT or Virgin line in the new place so I knew it might take a bit longer than usual to get setup. Also it would have been useful to have a connection while getting things sorted out, so I could work while waiting in for workmen. As stated at the start I’ve been pro Virgin in the past, I had their service at the old place and there was a CableTel (the Belfast cable company NTL acquired) access hatch at the property border so it was clear it had had service in the past. So on October 31st I placed an order on their website and was able to select an installation date of November 11th (earlier dates were available but this was a Saturday and more convenient).

This all seemed fine; Virgin contacted me to let me know there was some external work that needed to be done but told me it would happen in time. This ended up scheduled for November 9th, when I happened to be present. The engineers turned up, had a look around and then told me there was an issue with access to their equipment - they needed to do a new cable pull to the house and although the ducting was all there the access hatch for the other end was blocked by some construction work happening across the road. I’d had a call about this saying they’d be granted access from November 16th, so the November 11th install date was pushed out to November 25th. Unfortunate, but understandable. The engineers also told me that what would happen is the external team would get a cable hooked up to a box on the exterior of the house ready for the internal install, and that I didn’t need to be around for that.

November 25th arrived. There was no external box, so I was dubious things were actually going to go ahead, but I figured there was a chance the external + internal teams would turn up together and get it sorted. No such luck. The guy who was supposed to do the internal setup turned up, noticed the lack of an external box and informed me he couldn’t do anything without that. As I’d expected. I waited a few days to hear from Virgin and heard nothing, so I rang them and was told the installation had moved to December 6th and the external bit would be done before that - I can’t remember the exact date quoted but I rang a couple of times before the 6th and was told it would happen that day “for sure” each time.

December 5th arrives and I get an email telling me the installation has moved to December 21st. This is after the planned move date and dangerously close to Christmas - I’m aware that in the event of any more delays I’m unlikely to get service until the New Year. Lo and behold on December 7th I’m informed my install is on hold and someone will contact me within 5 working days to give me an update.

Suffice to say I do not get called. I ring towards the end of the following week and am told they are continuing to have trouble carrying out work on the access hatch. So I email the housing company doing the work across the road, asking if Virgin have actually been in touch and when the building contractors plan to give them the access they require. I get a polite response saying Virgin have been on-site but did not ask for anything to be moved or make it clear they were trying to connect a customer. And providing an email address for the appropriate person in the construction company to arrange access.

I contact Virgin to ask about this on December 20th. There’s no update but this time I manage to get someone who actually seems to want to help, rather than just telling me it’s being done today or soon. I get email confirmation that the matter is being escalated to the Area Field Manager (I’d been told this by phone on December 16th as well but obviously nothing had happened), and provide the contact details for the construction company.

And then I wait. I’m aware things wind down over the Christmas period, so I’m not expecting an install before the New Year, but I did think I might at least get a call or email with an update. Nothing. My wife rang to finally cancel our old connection last week (it’s been handy to still have my backup host online and be able to go and do updates in the old place) and they were aware of the fact we were trying to get a new connection and that there had been issues, but had no update and then proceeded to charge a disconnection fee, even though Virgin state no disconnection if you move and continue with Virgin Media.

So last week I rang and cancelled the order. And got the usual story of difficulty with access and asked to give them 48 hours to get back to me. I said no, that the customer service so far had been appalling and to cancel anyway. Which I’m informed has now been done.

Let’s be clear on what I have issue with here. While the long delay is annoying I don’t hold Virgin entirely responsible - there is construction work going on and things slow down over Christmas (though the order was placed long enough beforehand that this really shouldn’t have impacted things). The problem is the customer service and complete lack of any indication Virgin are managing this process well internally - the fact the interior install team turned up when the exterior work hadn’t been completed is shocking! If Virgin had told me at the start (or once they’d done the first actual physical visit to the site and seen the situation) that there was going to be a delay and then been able to provide a firm date, I’d have been much more accepting. Instead, the numerous reschedules, an inability to call back and provide updates when promised and the empty assurances that exterior work will be carried out on certain dates all leave me lacking any faith in what Virgin tell me. Equally, the fact they have charged a disconnection fee when their terms state they wouldn’t is ridiculous (a complaint has been raised but at the time of writing the complaints team has, surprise, surprise, not been in contact). If they’re so poor when I’m trying to sign up as a new customer, why should I have any faith in their ability to provide decent support when I actually have their service?

It’s also useful to contrast my Virgin experience with 2 others. Firstly, EE who I used for 4G MiFi access. Worked out of the box, and when I rang to cancel (because I no longer need it) were quick and efficient about processing the cancellation and understood that I’d been pleased with the service but no longer needed it, so didn’t do a hard retention sell.

Secondly, I’ve ended up with FTTC over a BT Openreach line from a local Gamma reseller, MCL Services. I placed this order on December 8th, after Virgin had put my install on hold. At the point of order I had an install date of December 19th, but within 3 hours it was delayed until January 3rd. At this point I thought I was going to have similar issues, so decided to leave both orders open to see which managed to complete first. I double-checked with MCL on January 2nd that there’d been no updates, and they confirmed it was all still on and very unlikely to change. And, sure enough, on the morning of January 3rd a BT engineer turned up after having called to give a rough ETA. Did a look around, saw the job was bigger than expected and then, instead of fobbing me off, got the job done. Which involved needing a colleague to turn up to help, running a new cable from a pole around the corner to an adjacent property and then along the wall, and installing the master socket exactly where suited me best. In miserable weather.

What did these have in common that Virgin does not? First, communication. EE were able to sort out my cancellation easily, at a time that suited me (after 7pm, when I’d got home from work and put dinner on). MCL provided all the installation details for my FTTC after I’d ordered, and let me know about the change in date as soon as BT had informed them (it helps I can email them and actually get someone who can help, rather than having to phone and wait on hold for someone who can’t). BT turned up and discovered problems and worked out how to solve them, rather than abandoning the work - while I’ve had nothing but good experiences with Virgin’s engineers up to this point there’s something wrong if they can’t sort out access to their network in 2 months. What if I’d been an existing customer with broken service?

This is a longer post than normal, and no one probably cares, but I like to think that someone in Virgin might read it and understand where my frustrations throughout this process have come from. And perhaps improve things, though I suspect that’s expecting too much and the loss of a single customer doesn’t really mean a lot to them.

Life’s been busy for the past few months, so excuse the lack of posts. One reason for this is that I’ve moved house. Virgin were supposed to install a cable modem on November 11th, but at the time of writing currently have my install on hold (more on that in the future). As a result when we actually moved in mid-December there was no broadband available. I’d noticed Currys were doing a deal on an EE 4GEE WiFi Mini - £4.99 for the device and then £12.50/month for 20GB on a 30 day rolling contract. Seemed like a good stopgap measure even if it wasn’t going to be enough for streaming video. I was pleasantly surprised to find it supported IPv6 out of the box - all clients get a globally routed IPv6 address (though it’s firewalled so you can’t connect back in; I guess this makes sense but it would be nice to be able to allow things through). EE are also making use of DNS64 + NAT64, falling back to standard CGNAT when the devices don’t support that.

All well and good, but part of the problem in the new place is a general lack of mobile reception in some rooms (foil backed internal insulation doesn’t help). So the MiFi is at the top of the house, where it gets a couple of bars of 4G reception and sits directly above the living room and bedroom. Coverage in those rooms is fine, but the kitchen is at the back of the house through a couple of solid brick walls and the WiFi ends up extremely weak there. Additionally my Honor 7 struggles to get a 3 signal in that room (my aging Nexus 7, also on 3, does fine, so it seems more likely to be the Honor 7 at fault here). I’ve been busy with various other bits over the Christmas period, but with broadband hopefully arriving in the new year I decided it was time to sort out my UniFi to handle coverage in the kitchen.

The long term plan is cabling around the house, but that turned out to be harder than expected (chipboard flooring and existing cabling not being in conduit ruled out the easy options, so there needs to be an external run from the top to the bottom). There is a meter/boiler room which is reasonably central and thus a logical place for cable termination and an access point to live. So I mounted the UniFi there, on the wall closest to the kitchen. Now I needed to get it connected to the MiFi, which was still upstairs. Luckily I have a couple of PowerLine adaptors I was using at the old place, so those provided a network link between the locations. The only remaining problem was that the 4GEE doesn’t have ethernet. What it does have is USB, and it presents as a USB RNDIS network interface. I had a spare DGN3500 lying around, so I upgraded it to the latest LEDE, installed kmod-usb-net-rndis and usb-modeswitch and then had a usb0 network device. I bridged this with eth0.1 - I want clients to talk to the 4GEE DHCP server so they can roam between the 2 APs, and I want the IPv6 configuration to work on both APs as well. I did have to change the IP on the DGN3500 as well - it defaulted to 192.168.1.1 which is what the 4GEE uses. Switching it to a static 192.168.1.2 ensures I can still get to it when the 4GEE isn’t active and prevents conflicts.

The whole thing ends up looking like the following (I fought Inkscape + Dia for a bit, but ASCII art turned out to be the easiest option):

/----------\       +-------+       +--------------+            +------------+
| Internet |--LTE--| EE 4G |--USB--|   DGN3500    |--Ethernet--| TL-PA9020P |
|          |       | MiFi  |       | LEDE 17.01.4 |            | PowerLine  |
\----------/       +-------+       +--------------+            +------------+
                       |                                              |
                      WiFi                                            |
                       |                                              |
                  +---------+                                         |
                  | Clients |                                     Ring Main
                  +---------+                                         |
                       |                                              |
                      WiFi                                            |
                       |                                              |
                  +--------+            +----------+            +------------+
                  | UniFi  |--Ethernet--|   PoE    |--Ethernet--| TL-PA9020P |
                  | AC Pro |            | Injector |            | PowerLine  |
                  +--------+            +----------+            +------------+

It feels a bit overly twisted for use with just a 4G connection, but various bits will be reusable when broadband finally arrives.

Floating point is a pain. I know this. But I recently took over the sigrok packages in Debian and as part of updating to the latest libsigkrok4 library enabled the post compilation tests. Which promptly failed on i386. Some narrowing down of the problem leads to the following test case (which fails on both gcc-6 under Debian/Stretch and gcc-7 on Debian/Testing):

#include <inttypes.h>
#include <stdio.h>
#include <stdint.h>

int main(int argc, char *argv[])
{
        printf("%" PRIu64 "\n", (uint64_t)((1.034567) * (uint64_t)(1000000ULL)));
}

We expect to see 1.034567 printed out. On x86_64 we do:

$ arch
x86_64
$ gcc -Wall t.c -o t ; ./t
1034567

If we compile for 32-bit the result is also as expected:

$ gcc -Wall -m32 t.c -o t ; ./t
1034567

Where things get interesting is when we enable --std=c99:

$ gcc -Wall --std=c99 t.c -o t ; ./t
1034567
$ gcc -Wall -m32 --std=c99 t.c -o t ; ./t
1034566

What? It turns out all the cXX standards result in the last digit incorrectly being 6, while the gnuXX standards (gnu11 is apparently the default) result in the correct trailing 7. Is there some postfix I can add to the value to prevent the floating point truncation taking place? Or do I just have to accept this? It works fine on armel, so it’s not a simple 32/64 bit issue.

I started writing this while sitting in Stansted on my way home from the annual UK Debian BBQ. I’m finally home now, after a great weekend catching up with folk. It’s a good social event for a bunch of Debian folk, and I’m very grateful that Steve and Jo continue to make it happen. These days there are also a number of generous companies chipping in towards the cost of food and drink, so thanks also to Codethink and QvarnLabs AB for the food, Collabora and Mythic Beasts for the beer and Chris for the coffee. And Rob for chasing us all for contributions to cover the rest.

I was trying to remember when the first one of these I attended was; trawling through mail logs there was a Cambridge meetup that ended up at Steve’s old place in April 2001, and we’ve consistently had the summer BBQ since 2004, but I’m not clear on what happened in between. Nonetheless it’s become a fixture in the calendar for those of us in the UK (and a number of people from further afield who regularly turn up). We’ve become a bit more sedate, but it’s good to always see a few new faces, drink some good beer (yay Milton), eat a lot and have some good conversations. This year also managed to get me a SheevaPlug so I could investigate #837989 - a bug with OpenOCD not being able to talk to the device. Turned out to be a channel configuration error in the move to new style FTDI support, so I’ve got that fixed locally and pushed the one line fix upstream as well.

I upgraded my last major machine from Jessie to Stretch last week. That machine was the one running the most services, but I’d made notes while updating various others to ensure it went smoothly. Below are the things I noted along the way, both for my own reference and in case they are of use to anyone else.

• Roundcube with the sqlite3 backend stopped working after the upgrade; fix was to edit /etc/roundcube/debian-db-roundcube.php and change sqlite3:// to sqlite:// in the $config['db_dsnw'] line.
• Dovecot no longer supports SSLv2 so had to remove !SSLv2 from the ssl_protocols list in /etc/dovecot/conf.d/10-ssl.conf
• Duplicity now tries to do a mkdir so I had to change from the scp:// backend to the sftp:// backend in my backup scripts.
• Needed to add needs_root_rights=yes to /etc/X11/Xwrapper.config so Kodi systemd unit could still start it on a new VT. Need to figure out how to get this working without the need for root.
• Upgrading fail2ban would have been easier if I’d dropped my additions in /etc/fail2ban/jail.d/ rather than the master config. Fixed for next time.
• ejabberd continues to be a pain; I do wonder if it’s worth running an XMPP server these days. I certainly don’t end up using it to talk to people myself.
• Upgrading 1200+ packages takes a long time, even when the majority of them don’t have any questions to ask during the process.
• PostgreSQL upgrades have got so much easier. pg_upgradecluster 9.4 main chugged away but did exactly what I needed.

Other than those points things were pretty smooth. Nice work by all those involved!