When I make a highly concurrent request to bitcoin core, a new block appears and all my requests get blocked #29384

1. 
   Rodert commented at 6:12 am on February 5, 2024: none

   Is there an existing issue for this?

   • I have searched the existing issues

   Current behaviour

   When I make a highly concurrent request to bitcoin core, a new block appears and all my requests get blocked

   Expected behaviour

   ok

   Steps to reproduce

   When I make a highly concurrent request to bitcoin core, a new block appears and all my requests get blocked

   Relevant log output

   No response

   How did you obtain Bitcoin Core

   Package manager

   What version of Bitcoin Core are you using?

   v25.0

   Operating system and version

   ubuntu 22.04 LTS

   Machine specifications

   No response

2. 
   glozow commented at 8:55 am on February 5, 2024: member

   Is it blocked forever or until the block is finished processing? Can you provide more details about what your highly concurrent request is?
3. 
   fanquake commented at 9:43 am on February 5, 2024: member

   Package manager

   Which package manager?

   Can you provide any debug.log output?

4. 
   Rodert commented at 9:23 am on February 6, 2024: none

   Is it blocked forever or until the block is finished processing? Can you provide more details about what your highly concurrent request is?

   this . I used golang to initiate rpc getblock and so on for the http request node, 30 concurrent continuous calls. As you can see from the load monitoring of the server, every time a new block is synchronized, the program is stuck.

   i use bitcoin version bitcoin-25.0 by Linux *.0-91-generic [#101](/bitcoin-bitcoin/101/)-Ubuntu SMP Tue Nov 14 13:30:08 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux

5. 
   Rodert commented at 9:30 am on February 6, 2024: none

   Package manager

   Which package manager?

   Can you provide any debug.log output?

   i use bitcoin version bitcoin-25.0 by Linux *.0-91-generic [#101](/bitcoin-bitcoin/101/)-Ubuntu SMP Tue Nov 14 13:30:08 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux .

   Here are some error messages encountered, But I don’t know if it matters.

    02024-02-05T06:13:24Z BlockUntilSyncedToCurrentChain: txindex is catching up on block notifications
    12024-02-05T06:27:32Z Saw new header hash=0000000000000000000380aa9f106ee0d0eb3ae7e0710d9562663f0d65c25de1 height=829003
    22024-02-05T06:27:32Z [net] Saw new cmpctblock header hash=0000000000000000000380aa9f106ee0d0eb3ae7e0710d9562663f0d65c25de1 peer=15
    32024-02-05T06:27:38Z UpdateTip: new best=0000000000000000000380aa9f106ee0d0eb3ae7e0710d9562663f0d65c25de1 height=829003 version=0x22000000 log2_work=94.711169 tx=962396874 date='2024-02-05T06:26:35Z' progress=1.000000 cache=210.9MiB(1465958txo)
    42024-02-05T06:27:38Z BlockUntilSyncedToCurrentChain: txindex is catching up on block notifications
    52024-02-05T06:27:38Z BlockUntilSyncedToCurrentChain: txindex is catching up on block notifications
    62024-02-05T06:27:38Z BlockUntilSyncedToCurrentChain: txindex is catching up on block notifications
    72024-02-05T06:27:38Z BlockUntilSyncedToCurrentChain: txindex is catching up on block notifications
    82024-02-05T06:30:30Z WARNING: request rejected because http work queue depth exceeded, it can be increased with the -rpcworkqueue= setting
    92024-02-05T06:30:30Z WARNING: request rejected because http work queue depth exceeded, it can be increased with the -rpcworkqueue= setting
   102024-02-05T06:30:30Z WARNING: request rejected because http work queue depth exceeded, it can be increased with the -rpcworkqueue= setting
   112024-02-05T06:30:30Z WARNING: request rejected because http work queue depth exceeded, it can be increased with the -rpcworkqueue= setting
   122024-02-05T06:30:36Z WARNING: request rejected because http work queue depth exceeded, it can be increased with the -rpcworkqueue= setting
   132024-02-05T06:30:36Z WARNING: request rejected because http work queue depth exceeded, it can be increased with the -rpcworkqueue= setting
   142024-02-05T06:30:36Z WARNING: request rejected because http work queue depth exceeded, it can be increased with the -rpcworkqueue= setting
   152024-02-05T06:30:36Z WARNING: request rejected because http work queue depth exceeded, it can be increased with the -rpcworkqueue= setting
   162024-02-05T06:30:36Z WARNING: request rejected because http work queue depth exceeded, it can be increased with the -rpcworkqueue= setting
   

6. 
   Rodert commented at 9:32 am on February 6, 2024: none

   I’ve reduced the concurrency now, the problem is happening less often, but it’s not completely solved
7. 
   maflcko commented at 9:41 am on February 6, 2024: member

   -rpcthreads= should be at least as large as the number of concurrent requests you are doing, otherwise there won’t be any threads to service the query and the work-queue will eventually overflow.
8. maflcko added the label Questions and Help on Feb 6, 2024
9. maflcko added the label RPC/REST/ZMQ on Feb 6, 2024
10. 
    maflcko commented at 10:11 am on February 6, 2024: member

    Related: #29386?
11. 
    maflcko commented at 6:06 pm on February 8, 2024: member

    Closing for now. Let us know if this is still an issue after #29384 (comment)
12. maflcko closed this on Feb 8, 2024

    ---

Contributors
Rodert glozow fanquake maflcko

Labels
RPC/REST/ZMQ Questions and Help