[LUDOC-494] Clarify when setting lnet route hops is required for Lustre 2.12 and Lustre 2.14

• Type: Bug
• Resolution: Unresolved
• Priority: Minor
• Severity: 3
• Assignee: Serguei Smirnov
• Reporter: Olaf Faaland

Description

In a discussion on https://review.whamcloud.com/#/c/43127/, in response to:

Note that bit of code is requiring that hop count be set for some routes, when they did not need to be set before (in lustre 2.12)

Chris said:

"Yes, good point. I think there was always an implicit requirement that hop count be set for multi-hop routes if the avoid_asym_route_failure feature was enabled, but we should make that explicit."

However this isn't reflected in the manual or lnetctl(8). (yet)

 

Comments

[Gerrit Updater]

"Andreas Dilger <[e-mail]>" merged in patch https://review.whamcloud.com/c/doc/manual/+/44916/
Subject: LUDOC-494 lnet: clarify use of route hopcount
Project: doc/manual
Branch: master
Current Patch Set:
Commit: f7da09ba79b2522ca51d001c59ab1212d051309c

[Gerrit Updater]

"Serguei Smirnov <[e-mail]>" uploaded a new patch: https://review.whamcloud.com/44916
Subject: LUDOC-494 lnet: clarify use of route hopcount
Project: doc/manual
Branch: master
Current Patch Set: 1
Commit: 48588c2fcdc74e48caca530f7e38f3036143ea95

[Olaf Faaland]

Hi Serguei,

We'll be happy to review the patches.

Thanks

[Serguei Smirnov]

Hi,

I reviewed the current related documentation, listed below are the recommended changes:

The lnetctl section of the Lustre manual and lnetctl man page should be updated to mention that the hop count defaults to 1 if not specified when adding a route with lnetctl.

Also, the manual should be updated to clarify that "avoid_asym_route_failure" module parameter applies only to single-hop routers. 

Also, the following passage from 34.3.7. LNet Peer Health should be modified: 

"A router is considered down if any of its NIDs are down. For example, router X has three NIDs: Xnid1, Xnid2, and Xnid3. A client is connected to the router via Xnid1. The client has router checker enabled. The router checker periodically sends a ping to the router via Xnid1. The router responds to the ping with the status of each of its NIDs. In this case, it responds with Xnid1=up, Xnid2=up, Xnid3=down. If avoid_asym_router_failure==1, the router is considered down if any of its NIDs are down, so router X is considered down and will not be used for routing messages. If avoid_asym_router_failure==0, router X will continue to be used for routing messages."  

The above sounds incorrect to me now, because the router shouldn't be considered down unless it cannot reach remote net.

Thanks,

Serguei.

 

[Peter Jones]

Serguei

Could you please advise on what changes should be made to the manual here?

Thanks

Peter