YDB CLI import stuck

[kunga]

Steps to reproduce:

1. Create table

   DROP TABLE `wikipedia`;
   
   CREATE TABLE wikipedia (
       id Uint64 NOT NULL,
       title Utf8,
       text Utf8,
     url Utf8,
     wiki_id Uint32,
     views float,
     paragraph_id Uint32,
     langs Uint32,
     emb Utf8,
     embedding String,
       PRIMARY KEY (id)
   );
   

2. Get dataset

   https://proxy.sandbox.yandex-team.ru/8368260097

3. Import dataset

   ydb -v import file csv --path wikipedia --header wikipedia_embeddings_train.csv --timeout 30
   

https://paste.yandex-team.ru/04758dbd-3d60-4eab-bf09-83b8c295111b/text

The upload will stop on ~73% when a datashard is splitting

There are some logs about retries but nothing is happen in the next 10 minutes

Also it may be noticed that there is no synchronization among threads on writing errors and the output is confusing (we have Sending retry attempt 1 of 10000 log 23 times

Maybe retry strategy uses exponential backoff and increases too much (x2 by each error in each thread), or there is some race and retries just dead locked

I propose that this issue must have a high priority as any new user that wants to try YDB may face it and never go back

[kunga]

Tried with --threads 1, also doesn't work and output is strange

https://paste.yandex-team.ru/0d69f59d-9486-48c6-b0be-79f4798eb799