[LU-10803] req_buffers_max and req_history_max setting problems

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.11.0
• Labels: patch
• Severity: 3
• Assignee: Bruno Faccini (Inactive)
• Reporter: Wang Shilong (Inactive)

Description

We hit LU-9372 OOM problems, and after applying

    LU-9372 ptlrpc: allow to limit number of service's rqbds

    we found two problems:

    

    1)Since 0 is a reserved value for @srv_nrqbds_max which

    means unlimited value, procfs write interface should support

    this value, otherwise, there is no way to change default behavior

    back.

 

    2)the check in ptlrpc_lprocfs_req_history_max_seq_write() was broken

    after this patch, the following check will always succeed if @srv_nrqbds_max

    is kept as default value 0:

    

    val > svc->srv_nrqbds_max/2

 

Issue Links

• is related to: LU-9372 OOM happens on OSS during Lustre recovery for more than 5000 clients

Comments

[Gerrit Updater]

Minh Diep ([e-mail]) uploaded a new patch: https://review.whamcloud.com/31665
Subject: LU-10803 ptlrpc: fix req_buffers_max and req_history_max setting
Project: fs/lustre-release
Branch: b2_10
Current Patch Set: 1
Commit: 14b6ab137329484507ddcb651167ce9b3ab5284a

[Peter Jones]

Landed for 2.11

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/31622/
Subject: LU-10803 ptlrpc: fix req_buffers_max and req_history_max setting
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: c11c61a82d668b54a41127a759b2e99cb077ac14

[Bruno Faccini (Inactive)]

Nice catch! And thanks for the patch.

 

[Wang Shilong (Inactive)]

Here is patch:

 

https://review.whamcloud.com/#/c/31622/