When commands fail to connect, they print additional error message "cannot set no_delay" and a different error code

Description

Introduced by the fix for PP-451.
We call set_nodelay after the clients connect to the server. However, if they fail also, we attempt to set_nodelay on a non-existent connection and that causes an extra, possibly confusing error message on the command line.

Reproduction
==========
Stop PBS server, and execute qstat.

[subhas@centosvm pbspro-private]$ qstat
No route to host
Cannot set nodelay on connection -1 (errno=113)
qstat: Server returned error 15033 for job

Acceptance Criteria

Test that when commands fail the additional error message is not printed.

Activity

Show:
Subhasis Bhattacharya
September 14, 2016, 6:50 AM

The fix is a single line change. A quick manual invocation of one of the commands can be the verification.

Subhasis Bhattacharya
September 14, 2016, 6:56 AM

After the single line fix, the output is as follows (and as it should be):

[subhas@centosvm pbspro]$ src/cmds/qstat
No route to host
qstat: cannot connect to server centosvm (errno=113)

Subhasis Bhattacharya
September 14, 2016, 6:57 AM

This was a simple miss introduced in and so any additional PTL test case would not be of any value. Its not a recurring issue and is easy to test manually, and needs to be tested only once.

Assignee

Subhasis Bhattacharya

Reporter

Subhasis Bhattacharya

Severity

None

OS

None

Start Date

None

Pull Request URL

None

Story Points

1

Components

Fix versions

Priority

Low
Configure