relay.py should not throw on socket exceptions, return False instead

RESOLVED FIXED

Status

Release Engineering
General
RESOLVED FIXED
6 years ago
2 months ago

People

(Reporter: Callek, Assigned: Callek)

Tracking

Firefox Tracking Flags

(Not tracked)

Details

Attachments

(1 attachment)

(Assignee)

Description

6 years ago
Created attachment 696538 [details] [diff] [review]
[tools] v1

I noticed in manual recovery today that relay.py is still not properly letting us retry, and is in fact throwing exceptions, which might be a cause of some of the panda dropouts and associated oranges.

This causes us to actually retry our 15 times as code says
Attachment #696538 - Flags: review?(jmaher)
(Assignee)

Comment 1

6 years ago
This should actually fix Bug 822321
Blocks: 822321
Comment on attachment 696538 [details] [diff] [review]
[tools] v1

Review of attachment 696538 [details] [diff] [review]:
-----------------------------------------------------------------

nice error handling procedure.  Can we mention it to the lifeguard folks as well, they are probably using the same code.
Attachment #696538 - Flags: review?(jmaher) → review+
(Assignee)

Comment 3

6 years ago
(In reply to Joel Maher (:jmaher) from comment #2)
> Comment on attachment 696538 [details] [diff] [review]
> [tools] v1
> 
> Review of attachment 696538 [details] [diff] [review]:
> -----------------------------------------------------------------
> 
> nice error handling procedure.  Can we mention it to the lifeguard folks as
> well, they are probably using the same code.

:dustin, remind me does lifeguard have something like this already, secondly where is the lifeguard version of relay.py now?
(In reply to Justin Wood (:Callek) from comment #3)
> :dustin, remind me does lifeguard have something like this already, secondly

Yes

> where is the lifeguard version of relay.py now?

mozpool/bmm/relay.py
(Assignee)

Comment 6

6 years ago
deployed.

and mozpool/mozpool/bmm/relay.py indeed has diverged a lot, and solved this by knowing it (in the same python instance) is all that will be controlling the relay host, in our case it can be multiple procs, even multiple systems. So I'm not worried beyond the "divergent code" worry. (but in theory we'll be moving to using mozpool for this soonish anyway)
(Assignee)

Updated

6 years ago
Status: NEW → RESOLVED
Last Resolved: 6 years ago
Resolution: --- → FIXED
Product: mozilla.org → Release Engineering
Component: General Automation → General
Product: Release Engineering → Release Engineering
You need to log in before you can comment on or make changes to this bug.