[MAPREDUCE-7329] HadoopPipes task may fail when linux kernel version change from 3.x to 4.x

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.6.0
• Fix Version/s: 3.4.0, 3.3.1
• Component/s: pipes
• Labels: patch pull-request-available
• Hadoop Flags: Reviewed
• Assignee: chaoli
• Reporter: chaoli
• Created: 15/Mar/21 06:42
• Updated: 09/Apr/21 03:11
• Resolved: 09/Apr/21 03:11

Description

Hadoop Pipes Ping implement has a bug. Recently, we upgrade linux kernel version from 3.x to 4.x. And we find hadoop pipe task exit with connect timeout which is implemented by PingThread in HadoopPipes.cc.

After a deep research, we finally find that current ping server won't accept ping client created socket, which may cause critical problem: 

•  it will cause tcp accept queue full(default 50)
•  when client close socket, server socket won't call close method, which will leave too many CLOSE_WAIT socket fd existed(default 2h), and accept queue never cleared.
• Even worse, in 4.x linux kernel version, it will cause tcp drop packet directly which makes ping client connect time out. While In 3.x linux kernel version, when accept queue full, client can also make half connection till sync queue full (default 2048), so from client side, ping will aslo work till sync queue full. And after 3 hours, task will also exit with connect timeout exception.

To fix this bug, we introduced a PingSocketCleaner thread, which will continuously accept ping socket connect from ping client. When socket close from client,  cleaner thread will detecte closed inputStream reading, then finally close socket from sever side.

Refrenced by linux kernel patch: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=5ea8ea2cb7

 

Issue Links

• links to: GitHub Pull Request #2775