Skip to main content

Hit by ransomware? Isolate affected systems now. Do not reboot or reformat.

SheMo Noransom舍末无勒

Case study

Sorry encrypts a cross-border e-commerce operator's cPanel servers

Sorry encrypted several cPanel / WHM servers through an authentication bypass, taking down site files and MySQL for more than a dozen storefronts with no public decryptor available. The cluster was rebuilt from offsite backups, a read replica at a second cloud provider, primary-side binlogs, CDN static copies and ERP order data.

Industry
Retail and E-commerce Ransomware Response
Family
Sorry
Scenario
MySQL Encrypted
Handled
2026-06

This case study is composed from typical scenarios to illustrate our approach and the recovery paths involved. It does not describe a specific customer incident; real cases involving customer data are published only with authorization.

  • 约 93%

    Site and business data recovered

  • 约 5 个工作日

    Storefronts back online

  • 9 台

    Servers rebuilt

  • 0

    Ransom paid

Incident background

The client is a cross-border e-commerce operator running a cluster of independent storefronts. More than a dozen sites serving European, North American and Southeast Asian markets sat on several Linux servers in one data centre, all managed through cPanel / WHM, with site files and each store's MySQL database on the same host.

To let an outsourced development team and agency staff maintain the sites, the WHM management port and SSH were exposed directly to the internet, and the same administrative password was reused across servers. Panel upgrades were scheduled quarterly by the hosting supplier, and the patch for CVE-2026-41940 had never entered a change window after release.

The incident began early on a weekend morning in June. On-call staff first saw monitoring alerts: SSH was unreachable on several servers and every MySQL process was offline. Customer service then reported database connection errors across all storefronts. Logging in from the out-of-band console, engineers found files in both the web and data directories carrying .sorry appended to their complete original names, with a README.md written into every encrypted directory. The note was identical on every site, carried a single fixed Tox ID, and contained neither a per-victim code nor a payment address.

Backups were "panel-scheduled jobs written to a second partition on the same host, plus a weekly sync to offsite object storage". The on-host copies were encrypted with the servers; the offsite copy was four days old.

What we did

Containment instructions went out on the first call: keep the affected hosts powered but off both networks, do not reboot, reinstall or repair anything on the original disks, and block SSH between servers.

  • Evidence and imaging. Read-only disk images of every affected host, plus panel and system logs, web logs and egress flow records. The original README.md and .sorry samples were preserved, and all later work ran on copies.
  • Family confirmation and recoverability. The .sorry extension, a README.md with a fixed Tox ID in every directory, .sorry_ payloads under /tmp and outbound .onion traffic identified this wave's 64-bit Go Linux ELF. It encrypts whitelisted extensions end to end with no intact regions left, so page-level repair does not apply and no usable public decryptor exists; recovery was scoped to backups and log replay.
  • Recovery execution. Site files came from the four-day-old offsite object storage copy; a read replica at a second cloud provider, outside the reach of this SSH spread, was the main database source, with unmatched binlogs on the primary replayed to bring the recovery point up to the encryption. Static assets came from the CDN and object storage, and orders were reconciled against the ERP / OMS copy.
  • Hardening. Compromised hosts were rebuilt rather than cleaned and cPanel / WHM was upgraded inside a standing patch window. The panel and SSH moved off the public internet behind a jump host with allowlisting and multi-factor authentication, SSH moved to keys, and shared passwords were eliminated. Backups were rebuilt to a three-copy design with offline and immutable copies.
  • Deliverables. A kill-chain timeline and compromised-host inventory, a statement of recovery scope and gaps, an exfiltration impact assessment with notification recommendations, and a hardening acceptance checklist. The report records both readings of the family's attribution - a new TellYouThePass variant to some vendors, a family new in 2026 to the national advisory.

Recovery result

All of the storefronts were publicly reachable again within roughly five business days, with product, order and member data recovered at about 93% overall. Order and payment records fared best, because the ERP / OMS held an independent copy. The gaps sat in recent campaign-page revisions, on-site reviews, user-uploaded media and two new storefronts not yet covered by replication; operations rebuilt those from the backup baseline and re-entered the rest by hand. The client paid no ransom and never contacted the operators through the Tox ID in the note.

Forensics confirmed that the attackers used the unpatched authentication bypass to gain administrative rights on the first server, then scanned internal SSH, brute-forced it with a built-in weak-password dictionary and rode the shared administrative password to the remaining hosts; nine servers required rebuilding. SSH was force-stopped and database processes killed before encryption began, which is why the alerts preceded any sign of changed extensions.

This family steals data and configuration files in bulk before encrypting, and no public leak site has been observed - which is not evidence that the data is safe. Egress flow data and web logs bounded the exfiltration window and the data types involved, including member phone numbers and delivery addresses. The client was advised to assess its notification duties under personal information protection requirements, inform its partner marketplaces and reset the affected passwords.

Follow-up recommendations: bring exposure and patch status into a monthly review; alert on anomalous panel authentication and bulk file renaming; raise offsite backups from weekly to daily; and retain logs for at least six months.

This is an illustrative case compiled from typical scenarios and anonymized; it does not refer to any specific client.

Emergency response

Data already encrypted? Stop and let an engineer look first

We do not pay ransoms and we do not negotiate with attackers. Engineers run a free assessment first, then propose a recovery plan and a firm quote.