Critical OpenVPN Core 3.x Failure After 1–3 Weeks – Config Corruption, Broken Core Toggle, Crash on Connect, and Urgent Need for v0.7.63

[CLONER-hub]

A severe stability issue affects OpenVPN for Android v0.7.62 on Android 13 (Samsung SM‑A127F, 64‑bit, India build). After 1–3 weeks of normal operation, all previously working VPN profiles stop functioning, configurations become corrupted, and the app ignores applied settings. Disabling the “OpenVPN Core 3.x” option does not work correctly—when the toggle is turned off and the user presses Connect, the application immediately crashes with no fallback and no diagnostic logs. The only recovery method is a full app data wipe.

The issue appears faster (within ~2 weeks) when using configuration files from Hidezen.com, after switching from a previously stable Hetzner Germany VPS. This indicates a Core 3.x state corruption or cache failure triggered by long‑running sessions or specific config structures.

Due to the severity of these failures and the inability to safely disable Core 3.x, there is an urgent need for the next optimized release (v0.7.63). Many users across Asia and North America are actively waiting for this update to restore long‑term stability and reliable Core 3.x behavior.

[schwabe]

You need to provide more details. This is the first report about something like that. And I don't have seen this in my own usage/testing.

Disabling the “OpenVPN Core 3.x” option does not work correctly—when the toggle is turned off and the user presses Connect, the application immediately crashes with no fallback and no diagnostic log

Please check if there is something in adb logcat.

Many users across Asia and North America are actively waiting for this update to restore long‑term stability and reliable Core 3.x behavior.

So far there has been only one report, yours.

[CLONER-hub]

I live in Iran, where the national firewall has been using very aggressive DPI against VPN protocols for many years (TLS fingerprint blocking, packet injection, MTU manipulation, TCP reset, throttling, etc.). Under these conditions, OpenVPN clients often behave very differently compared to clean networks in Europe or North America. In version 0.7.62 on my device, Core 3.x became very slow but stayed alive, while Core 2.x crashed immediately when I pressed Connect (no fallback and no internal diagnostic log). This was 100% reproducible under our local DPI and firewall conditions. Most users in Iran (and also many in Afghanistan) never file formal bug reports, even though they rely on tools like your client every day. Under censorship pressure, they keep changing configs, switching between many apps, and often move to other protocols such as Xray/XTLS/Reality instead of reporting issues. That is one reason why you might only see a single report like mine, even though the environment here is very hostile to OpenVPN. The important part is that after upgrading to 0.7.63pre from your server ( plai.de/android) and testing under the same conditions, Core 2.x no longer crashes and all my configs work normally. From my side, the crash I saw in 0.7.62 is now fixed in 0.7.63pre. If you still want a logcat from 0.7.62, I can reinstall that version and try to capture adb logs, but practically speaking the problem has been resolved for my environment. Thank you again for maintaining this project. Your client is still one of the most powerful OpenVPN clients available, and it makes a real difference for people like me living under heavy censorship. Sasan (CLONER‑hub) در تاریخ شنبه ۲۷ دسامبر ۲۰۲۵،‏ ۵:۴۹ ق.ظ Arne Schwabe < ***@***.***> نوشت:

…

*schwabe* left a comment (schwabe/ics-openvpn#1878) <#1878 (comment)> You need to provide more details. This is the first report about something like that. And I don't have seen this in my own usage/testing. Disabling the “OpenVPN Core 3.x” option does not work correctly—when the toggle is turned off and the user presses Connect, the application immediately crashes with no fallback and no diagnostic log Please check if there is something in adb logcat. Many users across Asia and North America are actively waiting for this update to restore long‑term stability and reliable Core 3.x behavior. So far there has been only one report, yours. — Reply to this email directly, view it on GitHub <#1878 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/BFDV6GQZ2LRY6EAL2BHRLPT4DXUCTAVCNFSM6AAAAACQCOCSWOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTMOJTGU3TQNJSG4> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[schwabe]

There is not really any change from 0.7.62 to that would explain that it got fixed for you. But without any stack traces or reproduction steps there is nothing I can do here.