RealTimeTrains website

[Tom]

We currently have a vacancy open for software engineers for the next few days and are currently interviewing. Ideally need a BSc in Computer Science (or higher), but I'm interested to hear from anyone with railway software experience. Please apply via https://www.realtimetrains.co.uk/about/jobs/ if you're interested - but you do need to able to interview this week. (We are potentially looking at taking two people on, I should point out.)

Tangentially related to RTT but I see Railmiles has gained a new field to record the type of journey. Leisure, business, crew or other. Nice idea but is there any chance of adding commuting as it feels different to business (at least to me) so would be nice to keep separate?

I can see where this new field is when entering a new journey but when viewing the same journey later I can’t see anything about it.

It's been a requested thing for a while and we have implemented it. It will become visible on public pages, with filtering, soon.

 

[ainsworth74]

Has been implemented

It's been a requested thing for a while and we have implemented it. It will become visible on public pages, with filtering, soon.

Excellent, thank you

 

[djpontrack]

It's been a requested thing for a while and we have implemented it. It will become visible on public pages, with filtering, soon.

Sounds interesting. Thank you Tom.

 

[SuperLuke2334]

TfW's 231s and 197s now have images showing with the allocation.

 

[Techniquest]

Ooh that's good news on the 231s, I saw the 197 ones early this morning. I wasn't expecting it, and initially I thought it was a 230 on a Manchester Airport working

 

[TT-ONR-NRN]

TfW's 231s and 197s now have images showing with the allocation.

Many thanks for this addition Tom

 

[kemal]

Morning

No train allocations shown on Real Time Trains so far today. Does anyone know what the problem is please?

Thanks Kemal

 

[Peter Mugridge]

Morning

No train allocations shown on Real Time Trains so far today. Does anyone know what the problem is please?

Thanks Kemal

It'll be the Network Rail data feeds playing up again.

 

[Davi]

It'll be the Network Rail data feeds playing up again.

Allocations are not part of the open data. Given the RTT status page shows everything working is normal, I'd say it's a problem with the operators.

 

[Tom]

This one looked like it was upstream of us but is in our control due to the fourth failure this week of separate components of our message queue infrastructure. We do have the data available so are currently working on reloading it.

In summary:

• Last Saturday night/Sunday night: disk space exhaustion due to durable queues not incorrectly being removed after a long time. This took out our London data centre fairly comprehensively.
• Tuesday morning: same issue but in our Manchester data centre which took London out with it for reasons we're still investigating. We suspect it's due to our deduplication strategy.
• Last night: we experienced a DoS attack on an unrelated part of our infrastructure which congested part of the switching behind the scenes.
• This morning: we discovered a split brain scenario in our London infrastructure where two MQ servers were acting as 'active' and processing different queues. That shouldn't happen so anything on the active server that shouldn't have been active... was routing direct to /dev/null. That was caused by a system misconfiguration that has not been picked up in 3 years and only occurred after a reboot happened a few days ago. How it didn't rear its head a few days ago I don't know.

A horrible week all in all from a systems perspective.

 

[RailWonderer]

Why are some 197s listed as having first class? Only some of the 3 car units will have first and none are in service yet. https://www.realtimetrains.co.uk/service/gb-nr:W26618/2023-10-13#allox_id=1

 

[Sheridan]

Why are some 197s listed as having first class? Only some of the 3 car units will have first and none are in service yet. https://www.realtimetrains.co.uk/service/gb-nr:W26618/2023-10-13#allox_id=1

Because it’s still partially planned for a MkIV set, the former Cardiff-Holyhead loco-hauled services are a bit messy in that way. Just needs amending.

 

[Tom]

1st class designations unless they're in the KYT diagrams are based on the planned schedule rather than the rolling stock operating.

 

[ainsworth74]

This one looked like it was upstream of us but is in our control due to the fourth failure this week of separate components of our message queue infrastructure. We do have the data available so are currently working on reloading it.

In summary:

• Last Saturday night/Sunday night: disk space exhaustion due to durable queues not incorrectly being removed after a long time. This took out our London data centre fairly comprehensively.
• Tuesday morning: same issue but in our Manchester data centre which took London out with it for reasons we're still investigating. We suspect it's due to our deduplication strategy.
• Last night: we experienced a DoS attack on an unrelated part of our infrastructure which congested part of the switching behind the scenes.
• This morning: we discovered a split brain scenario in our London infrastructure where two MQ servers were acting as 'active' and processing different queues. That shouldn't happen so anything on the active server that shouldn't have been active... was routing direct to /dev/null. That was caused by a system misconfiguration that has not been picked up in 3 years and only occurred after a reboot happened a few days ago. How it didn't rear its head a few days ago I don't know.

A horrible week all in all from a systems perspective.

Thank you for that comprehensive update! Sounds pretty awful. Is this also effecting Railmiles? Seems to be down this morning.

 

[Tom]

Nothing reported on our monitoring and it seems to be OK to me...

 

[ainsworth74]

Nothing reported on our monitoring and it seems to be OK to me...

Huh it does appear to be working. Very strange. Must have been a random glitch in my Internet connectivity or something. Sorry for duff report!

 

[Deepgreen]

RTT shows that the later in the day, the longer the trains seemingly are on the Berks and Hants line...!



RTT screenshot showing 9, 10, 15 and 18 car trains.

 

[JonathanH]

RTT shows that the later in the day, the longer the trains seemingly are on the Berks and Hants line...!

Logically, more reallocations happen later in the day.

 

[TT-ONR-NRN]

Site is down for me on both the phone and Mac.

Edit: Now seems to be fixed

 

[158841]

What are the chances of LNWR/WMT railway appearing? Getting hard to find the Reggie rail 323 and always end up with the same 350s! Help please tom?

 

[liamf656]

What are the chances of LNWR/WMT railway appearing? Getting hard to find the Reggie rail 323 and always end up with the same 350s! Help please tom?

If any operators are planned to show up on KYT it won't be announced until it actually appears because of the chance that the idea falls through last minute. Better to be pleasantly surprised than to wait for something that may not end up happening

 

[gazr]

The hovercraft sailings are not shown between Southsea and Ryde, where they appear under national rail planning...Is this by design? Searched SHV and "no trains found" (which is technically accurate )

 

[TT-ONR-NRN]

What are the chances of LNWR/WMT railway appearing? Getting hard to find the Reggie rail 323 and always end up with the same 350s! Help please tom?

Well I could just as easily plead for Greater Anglia. We all have individual TOCs that personally we want to see next. We’ll just have to wait and see. The TOCs have to approach RTT apparently, not the other way round.

 

[kbizzle]

Hello is there any further updates on the archived data? Would be great to fill in some back dates

 

[peteb]

Has anyone noticed that Real-time trains reports timings that don't tally with the actual time, be that BST or GMT? An example this morning is its reporting of 3S02 passing Kidderminster at 1151 1/2 when in fact the precise time was exactly 1150 GMT. Both my watch and phone showed 1150 and had been checked to the second with the 8am "pips". So why is a train recorded 1 1/2 minutes "later" than it really was, and how might this affect things like delay repay, assuming the data on RTT comes from the Network?

I have recently missed sightings of a few freight trains due to this anomaly.

 

[Lucy1501]

Has anyone noticed that Real-time trains reports timings that don't tally with the actual time, be that BST or GMT? An example this morning is its reporting of 3S02 passing Kidderminster at 1151 1/2 when in fact the precise time was exactly 1150 GMT. Both my watch and phone showed 1150 and had been checked to the second with the 8am "pips". So why is a train recorded 1 1/2 minutes "later" than it really was, and how might this affect things like delay repay, assuming the data on RTT comes from the Network?

I have recently missed sightings of a few freight trains due to this anomaly.

The data is recorded from the signalling system which works on its own timings. When a track circuit is occupied, the train description “steps” to the next berth and updates the timings.

In some locations, such as major stations, the signal is far back from the platform and it can take 1-2 minutes to actually get to the platform. The time recorded can therefore be out from the actual time it arrived, and public passenger timetables generally account for this.

It could also be that the train was travelling rather slow and the time between it passing the signal and the station could have been quite long.

If you want exact timings I’d recommend using something like Traksy. Traksy shows you the train describers in real time so you can see the exact positions of train at any time. The Kidderminster map is available here https://traksy.uk/live/M+31+KDRMNST.

 

[swt_passenger]

The data is recorded from the signalling system which works on its own timings. When a track circuit is occupied, the train description “steps” to the next berth and updates the timings.

In some locations, such as major stations, the signal is far back from the platform and it can take 1-2 minutes to actually get to the platform. The time recorded can therefore be out from the actual time it arrived, and public passenger timetables generally account for this.

It could also be that the train was travelling rather slow and the time between it passing the signal and the station could have been quite long.

If you want exact timings I’d recommend using something like Traksy. Traksy shows you the train describers in real time so you can see the exact positions of train at any time. The Kidderminster map is available here https://traksy.uk/live/M+31+KDRMNST.

To add to the above, the issue of what are known as “berth offsets” was explained a few years ago, Tom linked to his own blog post here:

RealTimeTrains website

Tom, One helpful upgrade would be for RTT to auto-refresh, at present whenever you look at a location, the results are shown at the time you press the search button and the page says the same until the user refreshes the page.

www.railforums.co.uk

AIUI fixing the timings in the way @Lucy1501 mentioned is more of a priority for passenger services?

 

[peteb]

The data is recorded from the signalling system which works on its own timings. When a track circuit is occupied, the train description “steps” to the next berth and updates the timings.

In some locations, such as major stations, the signal is far back from the platform and it can take 1-2 minutes to actually get to the platform. The time recorded can therefore be out from the actual time it arrived, and public passenger timetables generally account for this.

It could also be that the train was travelling rather slow and the time between it passing the signal and the station could have been quite long.

If you want exact timings I’d recommend using something like Traksy. Traksy shows you the train describers in real time so you can see the exact positions of train at any time. The Kidderminster map is available here https://traksy.uk/live/M+31+KDRMNST.

Thanks

 

[Tom]

I'm very surprised if 3S02 was at precisely 1150 given it was probably around 2km away from there at that point(!)

Its steps around that time were:

• 5840 -> 5838 at 114855
• 5838 -> 5834 at 115012
• 5834 -> 5822 at 115119
• 5822 -> 5820 at 115246

We have a Kidderminster arrival movement on 5838 5834 and a departure movement on 5834 5822.

Looking at the data coming through from that TD area now shows the times are essentially up to date so time drift isn't an issue. I've quickly looked at the signalling plan for that area and it looks about right to me.

I agree that 1151 1/2 is probably wrong - but given the way it calculates the passing time (averaging the difference between the "stopping" arrival and departure report) then I can see why its come to that. But it was probably at around 115110 by my reckoning.

 

[peteb]

I'm very surprised if 3S02 was at precisely 1150 given it was probably around 2km away from there at that point(!)

Its steps around that time were:

• 5840 -> 5838 at 114855
• 5838 -> 5834 at 115012
• 5834 -> 5822 at 115119
• 5822 -> 5820 at 115246

We have a Kidderminster arrival movement on 5838 5834 and a departure movement on 5834 5822.

Looking at the data coming through from that TD area now shows the times are essentially up to date so time drift isn't an issue. I've quickly looked at the signalling plan for that area and it looks about right to me.

I agree that 1151 1/2 is probably wrong - but given the way it calculates the passing time (averaging the difference between the "stopping" arrival and departure report) then I can see why its come to that. But it was probably at around 115110 by my reckoning.

Thanks for explanations etc. Very useful.