Open SimoneBosa opened 6 months ago
Hi,
just find my issue.
folder /etc/haproxy is a symbolic link to a path mounted on NFS server (over GFS). Haproxy configuration is shared between a cluster of a nodes.
Changing this mountpoint directly to GFS ( no more over NFS), it works without timeouts or errors.
Our suspects are on NFS server side (Ganesha) and inexplicable or unlikely locks on this cfg file.
Hope this can help others on same setup.
SB
Hi,
we are encountering the following issue when call dataplaneapi with curl.
When I call dataplane api with this curl
we got this message after 60 seconds
curl: (52) Empty reply from server
Looking at dataplane logs, we see this message
"POST /v2/services/haproxy/configuration/servers?backend=mybackend&version=1 HTTP/1.1\" 500 43 \"-\" \"curl/7.68.0\""
and we got a error 500.
Any idea about this strange behaviour?
Starting dataplane as a root we have
dataplaneapi -f /etc/dataplaneapi/dataplaneapi.yml
time="2024-03-13T17:05:05+01:00" level=info msg="Build date: 2024-02-15T08:20:47Z" time="2024-03-13T17:05:05+01:00" level=info msg="Build from: https://github.com/haproxytech/dataplaneapi" time="2024-03-13T17:05:05+01:00" level=info msg="HAProxy Data Plane API v2.9.1 4d10854c" time="2024-03-13T17:05:05+01:00" level=info msg="Reload strategy: systemd" time="2024-03-13T17:05:05+01:00" level=info msg="Serving data plane at http://[::]:5555"
/etc/dataplaneapi/dataplaneapi.yml
haproxy specs
haproxy.cfg include also