JR Ticket Machines and Ekinet Go Down Nationwide at Once: Why One Fault Stops All of JR — Inside the MARS System

Around late morning on July 1, 2026, reserved-seat ticket machines and the online booking services "Ekinet" and "e5489" across all of Japan's JR railway companies went down at once, making it hard to issue tickets. According to the rail-information account Trainfo, a fault occurred around 11:23 a.m. in "MARS," the core system that single-handedly runs JR's reservations and ticketing, and the impact spread to ticket machines and online services nationwide.

The striking question is why a single system's fault stops JR nationwide at the same time, from Hokkaido to Kyushu. MARS is the world's first seat-reservation system, born in 1960, and it still connects roughly 8,300 terminals across Japan and issues over 1.5 million tickets a day. This article first sorts out the situation and what travelers can do, then reads the structural weakness of "everything failing together" from the architecture of this huge system and the history of past outages — from the viewpoint of a working engineer.

*This is an ongoing incident. Times and scope are provisional, based on operator notices and reports; the cause and full-recovery time will be updated after official announcements. Note that a separate MobileSuica outage (charging suspended, etc.) also occurred on July 1, but that is a different system from MARS (see below).

What to do if you can't buy a ticket right now

First, the realistic actions for anyone about to travel. A MARS outage halts the "ticket-issuing" paths — machines and online booking — but tickets already in hand and some ticketless options tend to be less affected.

• Already-issued tickets and reserved seats can most likely still be used. If only issuing is blocked, ask station staff about handling after boarding.
• For same-day travel, switch to non-reserved seats or transit IC cards (Suica, ICOCA, etc.); you can keep moving even without a reserved ticket.
• If you absolutely need a reserved ticket, wait and retry the machine or Ekinet, or check the staffed ticket office. During an outage, offices use the same MARS, so they are not necessarily working either.
• Refunds and changes may not process correctly during the outage. Rather than repeating operations, it is safer to handle them after recovery. Ekinet's guidance for system troubles is also useful.

What makes this especially painful is stations with no nearby staffed ticket office. As we discuss later, JR East has sharply reduced its "Midori-no-Madoguchi" ticket offices in recent years, so when machines go down, travelers can be left with no one to ask.

What MARS is, and why the whole country stops at once

MARS (the passenger sales integrated system) is the core system that handles JR's reserved-seat and ticket booking and issuing on a nationwide scale. It is operated by Railway Information Systems Co. (JR Systems), founded in 1987 alongside the breakup and privatization of Japan National Railways and jointly owned by the seven JR companies. Its defining trait is that a single central system holds the inventory of "which seat on which train is free."

The scale is enormous. Per JR Systems' official materials, MARS connects about 8,300 terminals — roughly 5,600 at stations and ticket offices plus about 2,700 at travel agencies — and issues over 1.5 million tickets a day on average. At peak it handles 250 requests per second, returning a ticket in about six seconds on average. Its stated availability is 99.999% (a few minutes of downtime a year), an extremely high reliability target.

A common misunderstanding is that "Ekinet" and "e5489" are independent booking systems of their own — they are not. Ekinet (JR East) and e5489 (JR West) are only the front ends that users touch; the final seat-inventory check and reservation happen behind them by querying MARS's central system. So when MARS itself goes down, machines, ticket offices, Ekinet and e5489 all fail to issue tickets together — different entrances leading to the same place. The nationwide, all-at-once spread we saw is exactly this centralized structure showing through.

On the other hand, JR Central's EX-reservation and Smart-EX run on a system separate from MARS, even though they are also "online booking." In the 2020 outage discussed below, MARS-based credit-card payment stopped nationwide, yet Smart-EX and transit IC cards worked normally. Knowing "which services hang off MARS and which are separate" helps you judge whether your ticket is affected during an outage.

The world's first seat-reservation system, born 60 years ago

MARS's history is surprisingly old, and the fact that it is still in active service makes the story even more interesting. The first MARS-1 began running at Tokyo Station on February 1, 1960 — the world's first train seat-reservation system, handling reservations for the limited expresses "Tsubame" and "Hato." It was built by JNR's Railway Technical Research Institute and Hitachi. In an era before capable transistors, it used vacuum tubes alongside them, with a spinning magnetic drum for memory. The name "MARS" itself comes from this magnetic-drum-based automatic reservation origin. In 2025 the achievement was named an IEEE Milestone.

MARS then advanced generation by generation. MARS-101 in 1964 moved to full online real-time processing, and the next year it supported the Shinkansen and the new staffed ticket offices. MARS-301 in 1985 built the current backbone that unifies nationwide passenger sales; the 501 series from 2002 brought IP networking and a server-based design; and the 505 series in 2020 integrated ticketless and internet booking. Building on a 1960 design philosophy while swapping out its internals for over 60 years — this is the typical, grueling path a core system that underpins social infrastructure has to walk.

One fascinating point is that rail seat reservation is in some ways harder than airline booking. A plane manages "one segment, one seat," but rail must sell the same seat to different passengers by segment, like "Tokyo→Nagoya" and "Nagoya→Shin-Osaka." Since the magnetic-drum days, MARS has divided trains into segments and tracked each seat's usage to prevent double-selling — a hard inventory problem it has been solving all along. This seat-inventory logic, honed over half a century, is exactly the core asset of MARS that cannot easily be rebuilt.

"Everything failing together" has happened before

Nationwide outages are not new. Lining up past major failures of MARS and its surroundings, a pattern in the causes emerges.

The most symbolic is the February 2020 outage. Credit-card payment stopped nationwide at ticket offices, machines and online booking, and JR Systems disclosed the cause as "some fault occurred in a database attached to a server that makes up a subsystem." At that time, transit IC and JR Central's Smart-EX were unaffected — meaning only the parts hanging off MARS went down together. Rarely does the strength and weakness of a centralized design show so clearly.

The June 2023 large-scale JR East outage was caused by tripping a breaker that should not have been touched during power-equipment work — an error in the work procedure manual cut off server power. Beyond software bugs, a single physical point like power, and even human procedure, remain choke points where "if this one falls, everything falls." This is the reality of a giant core system.

From an engineer's view: the strength of centralization, and its price

Here I dig in a little from an engineer's viewpoint. MARS's design rests on the idea that "only one central system holds the correct seat inventory." Wherever a request comes from, the center checks inventory, reserves it, and authorizes the sale. This reliably prevents the same seat from being sold twice. For inventory that must never be oversold, like train seats, it is an extremely sound design.

The price is a single point of failure — the choke point where one broken part stops the whole. Because the center monopolizes the "truth" of inventory, if the center, or the communication, payment or power leading to it, falls anywhere, issuing across the country is affected at once. Designers knew this and layered defenses — duplication, a seismically isolated data center, a 99.999% availability target. Even so, from the "gaps in assumptions" — software updates, database faults, human error in power work, and outages of external services — a nationwide failure still happens every few years. The past cases sort cleanly into these four gap types.

So should it just go distributed? Not so simple. Splitting inventory across locations creates a consistency problem — "the same seat sold in two places at once" — and the cost of coordinating to prevent that soars. For reserved rail seats, where double-selling is unacceptable, centralization is both a "weakness" and a "rational choice." Herein lies the essential difficulty of rebuilding a core system that has run for 60 years. The struggles seen when companies try to move core systems onto new platforms echo the chaos of ANA's domestic system overhaul.

Why the impact is so large: the ticket-office cuts in the background

The same kind of outage hits society harder each year. One reason is that buying tickets has been pushed toward machines and online booking, while the staffed Midori-no-Madoguchi offices have been cut sharply. In May 2021, JR East announced a plan to reduce its offices from about 440 stations to roughly 140 — a cut of about 70%, anticipating a shift to online booking.

In reality, however, the shift online did not progress as expected, and lines at the offices grew worse. By April 2024 the cuts had reached 209 stations, but that May, JR East's president announced a "freeze" on the reduction plan and apologized, later shifting to reviving offices in busy seasons. In effect, the human fallback that was cut for efficiency is now being hastily restored.

This trend is the flip side of fragility during outages. The more you consolidate onto machines and online, the fewer escape routes remain when the MARS beneath them stops. When nationwide issuing halts during a daytime travel window as it did here, stations with few offices easily fall into a triple bind: "can't buy, can't ask, can only queue." Separate from the technical cause, an operational judgment — how much redundancy (backup means) to keep — shapes how large the damage feels to travelers.

The same-day MobileSuica outage is "a different story"

Because it is easily confused, let's separate it clearly. On the same July 1, MobileSuica also suffered an outage — in-app charging and the purchase of commuter and Green-car tickets stopped, and it entered emergency maintenance — but that is a problem in a different system from MARS. MobileSuica is JR East's e-money and ride-service platform, a different role from MARS, which handles seat reservations. Tap-to-ride with an existing balance and cash charging at convenience stores or station chargers are reportedly still possible.

Two JR-related outages happened to overlap on the same day, but the causes and targets are separate. When following the news, splitting "ticket issuing (MARS)" from "MobileSuica payment/charging" helps you correctly judge which one actually concerns you.

Summary

The nationwide ticketing trouble of July 1, 2026 was a fault in MARS, the core system that single-handedly runs JR's reservations, spreading at once to the machines, Ekinet and e5489 connected to it. The cause was under investigation at publication, but past cases show that from the "gaps in assumptions" — software updates, databases, power work, and external services — a nationwide outage has struck every few years. The world's first seat-reservation system, running since 1960, has centralized inventory to prevent double-selling; behind that rational design lives a structural weakness where one fallen point stops everything.

For travelers, the realistic move is to bridge with already-issued tickets or transit IC, avoid repeating operations, and wait for recovery. And this episode again poses a heavy question: how much of an outage "escape route" should a railway keep, after cutting the human counters in pursuit of convenience. We will update this article once an official cause is announced.

FAQ

References

• ・Railway Information Systems — MARS (passenger sales integrated system)
• ・Railway Information Systems — Company overview
• ・IPSJ Computer Museum — MARS-1
• ・IPSJ Computer Museum — MARS-101 and later
• ・Hitachi — MARS-1 named an IEEE Milestone (Aug 29, 2025)
• ・Nikkei xTECH — JR card-payment outage caused by a DB fault (Feb 2020)
• ・Nikkei xTECH — JR East's large outage from a mis-tripped breaker (Jun 2023)
• ・ITmedia — Why the Midori-no-Madoguchi cuts stalled
• ・Nikkei — JR East freezes ticket-office cuts (May 2024)
• ・Ekinet — Handling during system trouble