| Age | Commit message (Collapse) | Author |
|
Answering "is robw using this successfully, and the others?" from the
access logs took an hour of inference -- user agents, search terms, which
traits were mapped -- because nginx records the address and the subject is
only in the token. Two callers can share an address (an agent and a person
behind one NAT), one caller can move between addresses (an LLM's egress is
a /24), and the answer for a given person was "probably 80.115.1.132".
The API now sets `x-gn4-user: <token sub>` on every **authenticated**
response, and the `gn4` log format records it as `user=`. The subject is
only known inside the handler (the `AuthToken` extractor verifies the
token there), so a middleware opens a slot in the request extensions, the
extractor fills it, and the header is set on the way out; a request with no
valid token gets no header, which is the honest answer and reads as
`user=-`. The caller only ever sees an identity it already presented.
* src/auth.rs: `USER_HEADER`, `CallerSlot`, and the extractor recording
its subject.
* src/lib.rs: the `tag_caller` middleware, outermost so the slot exists
before any extractor runs.
* tests/auth.rs: `an_authenticated_response_names_its_caller` -- the header
equals the token's `sub`, and `/version` (no identity) carries none.
* etc/nginx-vesuvio.pu.ac.ke.conf (and the deployed copy of it): `user=`
in the log format.
* README.md, docs/nginx.md (the format, an `awk` one-liner for per-user
counts, and the caveat that old and new lines differ in field count),
docs/www/operator-checklist.md (`user=-` means no token, not a token
that failed late).
281 tests pass; the same four as before do not.
|
|
A caller posting a mapping result back to /plot/rqtl -- or a compute
request carrying a value per individual plus covariates -- was refused
with `413 Failed to buffer the request body: length limit exceeded`. That
is axum's DefaultBodyLimit, 2 MiB, which nothing here had raised: a real
GEMMA run on the BXD WGS panel is ~1.9 MB, so the plot endpoints sat just
under the cap and anything bigger (a genome-sized run, or every animal's
covariate value) fell over it. nginx's 10m was never the binding limit.
* src/lib.rs (BODY_LIMIT_BYTES): 64 MiB, with the reasoning and the note
that a proxy has to allow as much or it refuses first.
* etc/nginx-vesuvio.pu.ac.ke.conf, docs/nginx.md: client_max_body_size
64m, so the two limits agree instead of the proxy cutting in at 10m.
* README.md ("Request body size"): both caps, and that a request does not
need to be large -- `{"case_attribute": "Sex"}` has the server read the
attribute and align it, where posting a value per animal is pure payload.
Verified: the 11.9 MB body that was refused now returns an SVG; 67 MB
still gets axum's 413.
|
|
R/qtl and GEMMA are part of the service now, so the blanket 403 goes:
the handlers already require an AuthToken, which is the whole gate
("token_only", not resource-checked). What that buys and costs is now
written down rather than implied -- any token holder can name
server-side paths for R and gemma to read, and the public-data-only
restriction on traits/datasets does not apply here at all.
The box also has to be able to run them: PATH must carry Rscript,
gemma-wrapper and gemma, *and* R_LIBS_SITE must point at the profile's
site-library or scripts/rqtl_wrapper.R dies on `there is no package
called 'optparse'` -- which reads like a token or proxy fault and is
neither.
Verified through nginx with a token: POST /gn4/compute/rqtl 200 in 0.4s
with the fixture's marker rows, POST /gn4/compute/gemma 200 in 2.1s with
7 320 marker rows (identical to the documented run's first row),
GET /gn4/compute/gemma/status/<id> ending in "GN4: computation
complete", /compute/correlation 200; all four 401 without a token.
* etc/nginx-vesuvio.pu.ac.ke.conf: drop the /compute 403 locations, keep
them commented as the way back; timeouts/body-size comments now apply.
* docs/www/ai-instruction.md: new ยง4a, "Running QTL mapping (compute)" --
shapes, server-side paths, ?id= + status polling, what to do with a 504.
* docs/www/operator-checklist.md: compute moves from "must be off" to
"proxied, know what it grants", plus what the box needs (PATH,
R_LIBS_SITE, ravanan) and the measured results.
* docs/nginx.md: same, in the deployment and checking sections.
|
|
ai-instruction.md was written before anything was deployed: it took a
placeholder base URL, did not say where the API is mounted, and buried
the operator's deployment notes in an appendix labelled "not for the AI
reader" -- awkward to hand to an assistant, and confusing when it is.
So: the AI-facing page now names the deployment
(https://vesuvio.pu.ac.ke/gn4) and is self-contained, and the operator
material moves to operator-checklist.md. Everything in the page was
exercised against the running deployment.
Content changes, beyond the move:
* `next` hrefs arrive already prefixed (`/gn4/groups/BXD/strains`), so
they resolve against the origin -- joining them to the current URL's
directory double-prefixes them. Said plainly, since it is the one
thing the prefix breaks.
* says what cannot work and should not be retried: get-test-token is a
404, /compute/* is a 403 from the front end, and neither is a
permissions problem to work around.
* `/publications/{pubmed_id}/traits` wants the PubMed id, not the
internal `publication_id` in the same trait record, which 404s.
* states the current rate-limit position (none configured) instead of
implying a 429 policy that does not exist yet.
* docs/www/ai-instruction.md: rewritten, paste-ready.
* docs/www/operator-checklist.md: new file, the former appendix, updated
with what is now enforced at the proxy and a record of the verified
request matrix.
* README.md, docs/nginx.md, etc/nginx-vesuvio.pu.ac.ke.conf: point at
the operator checklist by its new name.
|
|
Without it /search forwards to GeneNetwork3 at GN3_URL (:5003) and
answers 502 backend_unreachable when nothing is listening there -- a
correct error, with the error body's links rewritten, but easy to
misread as the proxy being wrong. The deployment has an index, so native
is what it wants.
* docs/nginx.md (The server behind nginx): add the variable and say what
the default does.
* etc/nginx-vesuvio.pu.ac.ke.conf: same, in the start-up comment.
|
|
Reproduced from a mirror of the deployed /etc/nginx: with
sites-enabled/default still present, nginx logs
conflicting server name "vesuvio.pu.ac.ke" on 0.0.0.0:443, ignored
and answers 443 with the default site -- stock welcome page at /,
nginx's own 404 for /gn4/..., HTTP/1.1 despite `http2 on` in this block.
sites-enabled/* is globbed alphabetically and the first block whose
server_name matches wins, so `default` (edited by certbot to serve this
hostname) shadows the gn4 block. Removing it gives the expected 302 /
301 / 200 / 404 / 403 matrix.
* docs/nginx.md (nginx): say to install the snippet instead of the
distribution default, and what the shadowing looks like from outside.
|
|
The certificate on :443 carries subjectAltName DNS:vesuvio.pu.ac.ke, and
vesuvio.pu.ac.uk does not resolve at all, so the config as committed
could not have matched a request for the real name: nginx falls through
to the default server, which is why /gn4/... answered 404 from nginx
rather than a 502 from genenetwork4.
* etc/nginx-vesuvio.pu.ac.uk.conf -> etc/nginx-vesuvio.pu.ac.ke.conf:
server_name and the certificate paths.
* docs/nginx.md: the same rename, and a "Checking it" section that
starts from what actually gets in the way on the box -- TLS validates
the URL hostname rather than a Host: header, so on-box checks need
--resolve against the name on the certificate, and a mismatch there and
in server_name look identical (default server answers).
|
|
docs/nginx.md explained the pieces; this is the assembled file for the
deployment actually being set up, so it can be reviewed and installed
without stitching anything together. It lives in etc/ rather than
docs/, being deployable configuration rather than prose.
It refuses /compute/*, which the README-based template did not: those
endpoints take server-side file paths and run gemma/ravanan/R on them,
so any token holder able to reach them can read the box -- the operator
checklist in docs/www/ai-instruction.md already said to keep them on the
internal listener, and this is that rule where it belongs.
* etc/nginx-vesuvio.pu.ac.uk.conf: new file. TLS, http2, log format
carrying x-request-id, / and /gn4 redirects, proxy_pass with the
prefix-stripping trailing slash, raised timeouts, sub_filter rewriting
of href/next/spec_url/type, and 403/404 refusals for /compute/ and
/get-test-token.
* docs/nginx.md: point at it, and say why /compute/ is refused.
|
|
The deployment being set up is https://vesuvio.pu.ac.uk/gn4 proxying to
a loopback port, with no gn-auth yet. Two things there are not obvious
and neither is discoverable from the README's /get-test-token section:
every link the API emits is root-relative, so under /gn4 they resolve
outside the mount point; and /compute/* runs for minutes in a single
request, so nginx's one-minute default read timeout cuts it off
mid-computation.
The link rewriting is the delicate part -- the server's serde output is
compact ("href":"/"), not the pretty-printed shape the README examples
show, so a filter written from those examples silently does nothing.
Both spellings of all four affected keys (href, next, spec_url, type)
are in the config, and the envelope's "next":[ array is deliberately
left alone.
* docs/nginx.md: new file. Server environment, the nginx block, what the
prefix does to links and the three ways to handle it, stub-minted
tokens instead of gn-auth, and how to check the result.
* README.md (Behind nginx or in a container): point at it.
|