Skip to content

Leading bytes dropped from first SSH channel data under concurrent connects #7

Description

@bdolenc

Under high concurrent SSH connect load, the first data delivered on a newly opened channel can be missing its leading bytes.

Observed with ~250 concurrent NETCONF subsystem sessions. A small fraction of sessions receive a corrupt server <hello> and fail XML parsing.

Example corrupt first message:

urn:ietf:params:xml:ns:yang:iana-if-type?module=iana-if-type&amp;revision=2014-05-08</capability><capability>urn:ietf:params:xml:ns:yang:ietf-restconf-monitoring?...

the <?xml...><hello><capabilities><capability> prefix is missing.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions