No saved hashing after crashing

[guglovich]

I was able to hache and save only one folder at a time. In other sessions, I do hashing of several folders, after some time the client crashes, and even the first folder for which I did hashing is not saved. Apparently, the client stores either everything or nothing. Or such data is stored in RAM. Anyway, I haven't been able to enlarge my collection for a long time.

Current UI and client versions:

2.14.0

Operating system:

Arch Linux x86_64
Firefox 145

Steps to reproduce the issue:

It happens every time you add hundreds of gigabytes to a new ball. I choose different folders, but the problem repeats anyway.

Screenshots

[maksis]

Please see https://github.com/airdcpp-web/airdcpp-webclient/blob/master/.github/CONTRIBUTING.md for crash reporting instructions

[maksis]

Or could it be a duplicate of this: airdcpp/airdcpp-windows#193 ?

[guglovich]

Time: 17:14:51
OS version: Linux 6.17.6-x64v3-xanmod1 (x86_64)
Client version: AirDC++w 2.14.0

airdcppd in function [unknown] [0x55e01d8d72ac]
/usr/lib/libc.so.6 in function [unknown] [0x7f0eb2b384d0]
/usr/lib/libairdcpp.so.2.14.0 in function dcpp::Client::onPrivateMessage(std::shared_ptrdcpp::ChatMessage const&) [0x7f0eb37353ed]
/usr/lib/libairdcpp.so.2.14.0 in function dcpp::NmdcHub::onLine(std::__cxx11::basic_string<char, std::char_traits, std::allocator > const&) [0x7f0eb3759be7]
/usr/lib/libairdcpp.so.2.14.0 in function dcpp::BufferedSocket::threadRead() [0x7f0eb367fb31]
/usr/lib/libairdcpp.so.2.14.0 in function dcpp::BufferedSocket::run() [0x7f0eb36814ff]
/usr/lib/libairdcpp.so.2.14.0 in function dcpp::Thread::starter(void*) [0x7f0eb36cb04e]
/usr/lib/libc.so.6 in function [unknown] [0x7f0eb2b9098b]
/usr/lib/libc.so.6 in function [unknown] [0x7f0eb2c149cc]

[maksis]

That's a cryptical crash. Could you paste the log also for another crash to see whether it looks the same?

[guglovich]

I have not been able to reproduce the problem right now. I assume that the crash occurs if hubs are connected and folders are hashed. This time, I hashed my huge folders without connecting to hubs.

I also want to point out that the second point of the error report is that the hash was not saved during such failures.

[maksis]

I also want to point out that the second point of the error report is that the hash was not saved during such failures.

It's saved in such cases. However, you haven't mentioned that how many files had completed hashing at the time of the crash (maybe there was nothing to save).

[guglovich]

I also want to point out that the second point of the error report is that the hash was not saved during such failures.

It's saved in such cases. However, you haven't mentioned that how many files had completed hashing at the time of the crash (maybe there was nothing to save).

It was over 200 GB, that's for sure. But it could have been 300+. After the crash, only the very first folder was saved, allocated to a separate hashing session.