[LU-13641] socklnd: remove use_tcp_bonding option in favor of LNet Multi-Rail

• Type: Technical task
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.16.0, Lustre 2.15.0
• Affects Version/s: Lustre 2.14.0, Lustre 2.12.4, Lustre 2.15.0
• Labels: lnet
• Assignee: Serguei Smirnov
• Reporter: Amir Shehata (Inactive)

Description

TCP bonding in socklnd over-complicates the code and there is no evidence it's being used anywhere. With LNet Multi-Rail, the use_tcp_bonding option has become obsolete. Add a deprecation message for earlier releases. Remove it in the 2.15 release.

Multi-Rail feature doesn't need to be explicitly enabled. To use MR instead of the use_tcp_bonding configuration option, group the interfaces on the same network using the lnetctl utility:

lnetctl net add --net tcp --if eth0,eth1

or via the /etc/modprobe.d/lnet.conf or /etc/modprobe.d/lustre.conf configuration file:

options lnet networks="tcp(eth0,eth1)"

and make sure dynamic discovery is enabled:

lnetctl set discovery 1

MR will aggregate the throughput of all available networks/interfaces shared between peer nodes.

See LNet Software Multi-Rail Configuration in the Lustre Operations Manual for more details.

Comments

[Gerrit Updater]

"Oleg Drokin <[e-mail]>" merged in patch https://review.whamcloud.com/c/fs/lustre-release/+/48568/
Subject: LU-13641 socklnd: remove remnants of tcp bonding
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 3630e1eaf9db562a1de707762cd649db815459c8

[Gerrit Updater]

"Neil Brown <[e-mail]>" uploaded a new patch: https://review.whamcloud.com/48568
Subject: LU-13641 socklnd: remove remnants of tcp bonding
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: 1f3fd1fd3dc9dc565fd7691316324b287d161148

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/40774/
Subject: LU-13641 socklnd: replace route construct
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 7766f01e891c378d1bf099e475f128ea612488f0

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/40000/
Subject: LU-13641 socklnd: remove tcp bonding
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: d123c47a18adbf5665ed63d99c53117b84db9ec8

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/41102/
Subject: LU-13641 socklnd: announce deprecation of 'use_tcp_bonding'
Project: fs/lustre-release
Branch: b2_12
Current Patch Set:
Commit: ede08af7d30b2dc4c41b89db224ab1a3bdb2f30c

[Gerrit Updater]

James Simmons ([e-mail]) uploaded a new patch: https://review.whamcloud.com/41180
Subject: LU-13641 socklnd: replace route construct
Project: fs/lustre-release
Branch: b2_12
Current Patch Set: 1
Commit: 5364ae3875337160e5928631dd7484703789ea47

[Gerrit Updater]

James Simmons ([e-mail]) uploaded a new patch: https://review.whamcloud.com/41179
Subject: LU-13641 socklnd: remove tcp bonding
Project: fs/lustre-release
Branch: b2_12
Current Patch Set: 1
Commit: 0124602e645e87e1e51e5b218a823f817e457547

[Peter Jones]

The deprecation warning has landed to 2.14. The removal itself is deferred to 2.15

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/41088/
Subject: LU-13641 socklnd: announce deprecation of 'use_tcp_bonding'
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 1a2bf911b9793648de3abbc88b9f77eb2237dc50

[Cory Spitz]

Ah, thanks for the clarification. That makes sense.

[Serguei Smirnov]

Hi Cory,

The wording is a bit confusing, so I'll clarify that this ticket is dealing with just a socklnd feature, so one would still be able to use native TCP bonding in Linux with MR. It is the "socklnd bonding" that's being deprecated. Introduction of "socklnd bonding" allowed treating multiple interfaces as one in socklnd layer - introduction of MR brought the same concept to LNet layer. I don't believe there's a difference in performance. ashehata can correct me if my understanding is wrong.