DPDK CI discussions
 help / color / mirror / Atom feed
From: Aaron Conole <aconole@redhat.com>
To: Michael Santana <msantana@redhat.com>
Cc: ci@dpdk.org,  Eelco Chaudron <echaudro@redhat.com>,
	 Ilya Maximets <i.maximets@ovn.org>,
	 Dumitru Ceara <dceara@redhat.com>
Subject: Re: [PATCH] cirrus: add an initial polling mechanism
Date: Fri, 19 Jan 2024 09:51:33 -0500	[thread overview]
Message-ID: <f7t7ck59y16.fsf@redhat.com> (raw)
In-Reply-To: <CABVNPRrpkhEo_6whkUXeasTbSTAL9420Bq4_-_+SSOJkZ4dAtw@mail.gmail.com> (Michael Santana's message of "Mon, 15 Jan 2024 15:00:48 -0500")

Michael Santana <msantana@redhat.com> writes:

> On Tue, Dec 12, 2023 at 11:08 AM Aaron Conole <aconole@redhat.com> wrote:
>>
>> Cirrus-CI is used to provide a testing environment for FreeBSD, among
>> others.  As of now, the ci monitoring framework doesn't support checking
>> for cirrus-ci status, but that can be changed.  We add an initial polling
>> similar to the original GitHub Actions or Travis-CI polling mechanisms
>> that merely provides pass/fail/warn with a URL.  This can be improved
>> later to provide logs on a per-task (similar to GHA Jobs) basis.
>>
>> Signed-off-by: Aaron Conole <aconole@redhat.com>
> For the most part it looks good to me. See inline comments below. Thanks Aaron!
>> ---
>>  ci_mon           |   6 ++-
>>  cirrus_mon       | 100 +++++++++++++++++++++++++++++++++++++++++++++++
>>  series_db_lib.sh |  13 +++++-
>>  3 files changed, 116 insertions(+), 3 deletions(-)
>>  create mode 100755 cirrus_mon
>>
>> diff --git a/ci_mon b/ci_mon
>> index 2e30da4..4b35d38 100755
>> --- a/ci_mon
>> +++ b/ci_mon
>> @@ -140,13 +140,16 @@ if [ "X" = "X$pw_instance" ]; then
>>     exit 1
>>  fi
>>
>> -for I in travis github obs dummy; do
>> +for I in travis github obs cirrus dummy; do
> ok
>>
>>      token=${I}_token
>>      disable=disable_${I}
>>
>>      if [ "X${!disable}" = "Xyes" ]; then
>> +        echo "Skiping ${I}"
>>          continue
>> +    else
>> +        echo "Scanning ${I}"
>>      fi
>>
>>      ./${I}_mon $pw_instance ${!token} "$pw_project" | grep "^pw|" | while IFS="|" \
>> @@ -163,6 +166,7 @@ for I in travis github obs dummy; do
>>
>>          patch_id="$(get_patch_id_by_series_id_and_sha "$series_id" "$shasum" "$pw_chk_instance")"
>>          if [ "X$patch_id" == "X" ]; then
>> +            echo "No patchid..."
>>              patch_id=$(echo $PATCHDATA | jq -rc '.patches[-1].id')
>>          fi
>>          PATCHDATA="$(echo "$PATCHDATA" | jq ".patches[] | select(.id==$patch_id)")"
>> diff --git a/cirrus_mon b/cirrus_mon
>> new file mode 100755
>> index 0000000..f307961
>> --- /dev/null
>> +++ b/cirrus_mon
>
> I agree with your commit message. This is very much like the github
> and travis pulling. A lot of stuff that's similar to those scripts.
>
> So most of my questions are based on the cirrus specific workflow. It
> has been a while since I have worked with cirrus

Okay

>> @@ -0,0 +1,100 @@
>> +#!/bin/bash
>> +# SPDX-Identifier: gpl-2.0-or-later
>> +# Copyright (C) 2023, Red Hat, Inc.
>> +#
>> +# Monitors cirrus build history for builds in a series.
>> +# Records the builds in the series database (and emits them on the
>> +# stdout line for processing)
>> +#
>> +# Licensed under the terms of the GNU General Public License as published
>> +# by the Free Software Foundation; either version 2 of the License, or
>> +# (at your option) any later version.  You may obtain a copy of the
>> +# license at
>> +#
>> +#    https://www.gnu.org/licenses/old-licenses/gpl-2.0.html
>> +#
>> +# Unless required by applicable law or agreed to in writing, software
>> +# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
>> +# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
>> +# License for the specific language governing permissions and limitations
>> +# under the License.
>> +
>> +[ -f "$(dirname $0)/series_db_lib.sh" ] && source "$(dirname $0)/series_db_lib.sh"
>> +
>> +while [ "$1" != "" ]; do
>> +    if echo "$1" | grep -q -s -E ^--pw-instance= ; then
>> +        pw_instance=$(echo "$1" | sed s/^--pw-instance=//)
>> +        shift
>> +    elif echo "$1" | grep -q -s -E ^--cirrus-token= ; then
>> +        cirrus_token=$(echo "$1" | sed s/^--cirrus-token=//)
>> +        shift
>> +    elif echo "$1" | grep -q -s -E ^--help ; then
>> +        echo "Cirrus CI monitor script"
>> +        echo "$0: args"
>> +        echo "Required if not set in ~/.pwmon-rc file:"
>> +        echo "  --pw-instance=<inst url>       URL for pw"
>> +        echo ""
>> +        echo "Options:"
>> +        echo "  --cirrus-token=<token>         Sets tokenfor web requests"
> s/tokenfor/token for/
>> +        echo ""
>> +        exit 0
>> +    elif ! echo "$1" | grep -q -s -E ^-- ; then
>> +        break
>> +    fi
>> +done
>> +
>> +echo "scanning started " >> /tmp/cirrus.log
>> +
>> +if [ "X" = "X$pw_instance" ]; then
>> +    pw_instance="$1"
>> +    shift
>> +fi
>> +
>> +if [ "X" = "X$github_token" ]; then
> How come cirrus needs a github token? Sorry, it's been a while since I
> used cirrus. Are cirrus builds based on github commits?

I'll double check - IIRC, cirrus uses the github token as a log in, but
it might be a bad variable name.

>> +    github_token="$1"
>> +    shift
>> +fi
>> +
>> +if [ "X" = "X$pw_project" -a "X$1" != "X" ]; then
>> +    pw_project="$1"
>> +    shift
>> +fi
>> +
>> +if [ "X$cirrus_token" != "X" ]; then
>> +    AUTH="-H \"Authorization: Bearer $cirrus_token\""
>> +fi
>> +
>> +ci_instance="cirrus_sync"
>> +
>> +get_unsynced_series "$pw_instance" "$ci_instance" | \
>> + while IFS="|" read -r series_id patch_id patch_url patch_name sha
>> patchwork_instance patchwork_project repo_name gap_sync obs_sync
>> cirrus_sync; do
>> +        repo_owner=$(echo "$repo_name" | cut -d/ -f1)
>> +        repo_real=$(echo "$repo_name" | cut -d/ -f2)
>> +
>> + graph_string="{ \"query\": \"query BuildBySHAQuery(\$owner:
>> String!, \$name: String!, \$SHA: String){
>> searchBuilds(repositoryOwner: \$owner, repositoryName: \$name, SHA:
>> \$SHA) { id, status } }\", \"variables\": { \"owner\":
>> \"$repo_owner\", \"name\": \"$repo_real\", \"SHA\": \"$sha\" } }"
>> + build_details=$(curl -s "$AUTH" -X POST --data "$graph_string"
>> https://api.cirrus-ci.com/graphql)
> My other question regarding cirrus workflow was this. Normally with
> github we would have a direct URL to the project, e.g.
> github.com/ovsrobot/ovs, but I dont see anything like
> "cirrus.com/ovsrobot/ovs". Maybe that inmbeded in one of the
> variables? or is this why we have the query string using the variables
> repo_owner and repo_real? If that's the case then it's a shame cirrus
> doesnt have a direct API call into the project itself

Cirrus doesn't do API calls like that.  Instead, we need to use GraphQL
to make requests.  The whole API is documented, but there are parts of
it that surely need more details than others.

Actually, the reason that this first pass doesn't break out all the jobs
is because the GraphQL queries have to get a bit complex to build that
detail up.  It makes the API very powerful, but also very difficult to
work with for simple stuff.

>> +
>> +        id=$(echo "$build_details" | jq -rc '.data.searchBuilds[-1].id')
>> +        status=$(echo "$build_details" | jq '.data.searchBuilds[-1].status')
>> +        build_url="https://cirrus-ci.com/build/$id"
>> +
>> +        result="in-progress"
>> +        if [ "$status" == "\"COMPLETED\"" ]; then
>> +            result="passed"
>> +        elif [ "$status" == "\"FAILED\"" ]; then
>> +            result="failed"
>> +        elif [ "$status" == "\"ABORTED\"" ]; then
>> +            set_synced_patch "$patch_id" "$patchwork_instance" "$ci_instance"
>> + echo "CIRRUS patch_id=$patch_id belonging to series=$series_id on
>> $patchwork_instance was aborted" 1>$2
>> +            continue
>> +        elif [ "$status" == "\"ERRORED\"" ]; then
>> +            result="warn"
>> +        fi
>> +
>> +        if [ "$result" == "in-progress" ]; then
>> + echo "CIRRUS patch_id=$patch_id belonging to series=$series_id is
>> not completed[$status]. Skipping" 1>&2
>> +            continue
>> +        fi
>> +
>> + echo
>> "pw|$pw_instance|build|$series_id|SHA|$sha|$result|$build_url|$patch_name|$repo_name|$test_name"
> ok
>> +done
>> diff --git a/series_db_lib.sh b/series_db_lib.sh
>> index c5f42e0..e415870 100644
>> --- a/series_db_lib.sh
>> +++ b/series_db_lib.sh
>> @@ -130,6 +130,15 @@ recheck_sync INTEGER
>>  EOF
>>          run_db_command "INSERT INTO series_schema_version(id) values (8);"
>>      fi
>> +
>> +    # 0009 - cirrus CI data
>> +    run_db_command "select * from series_schema_version;" | egrep '^9$' >/dev/null 2>&1
>> +    if [ $? -eq 1 ]; then
>> +        sqlite3 ${HOME}/.series-db <<EOF
>> +ALTER TABLE git_builds ADD COLUMN cirrus_sync INTEGER;
>> +EOF
>> +        run_db_command "INSERT INTO series_schema_version(id) values (9);"
> ok
>> +    fi
>>  }
>>
>>  function series_db_exists() {
>> @@ -379,7 +388,7 @@ function set_synced_for_series() {
>>
>>      series_db_exists
>>
>> -    echo "update git_builds set gap_sync=1, obs_sync=1 where patchwork_instance=\"$instance\" and series_id=$series_id;" | series_db_execute
>> +    echo "update git_builds set gap_sync=1, cirrus_sync=1, obs_sync=1 where patchwork_instance=\"$instance\" and series_id=$series_id;" | series_db_execute
> ok
>>  }
>>
>>  function set_unsynced_for_series() {
>> @@ -402,7 +411,7 @@ function insert_commit() {
>>
>>      series_db_exists
>>
>> -    echo "INSERT INTO git_builds (series_id, patch_id, patch_url, patch_name, sha, patchwork_instance, patchwork_project, repo_name, gap_sync, obs_sync) VALUES($series_id, $patch_id, \"$patch_url\", \"$patch_name\", \"$sha\", \"$instance\", \"$project\", \"$repo_name\", 0, 0);" | series_db_execute
>> +    echo "INSERT INTO git_builds (series_id, patch_id, patch_url, patch_name, sha, patchwork_instance, patchwork_project, repo_name, gap_sync, obs_sync, cirrus_sync) VALUES($series_id, $patch_id, \"$patch_url\", \"$patch_name\", \"$sha\", \"$instance\", \"$project\", \"$repo_name\", 0, 0, 0);" | series_db_execute
> ok
>>  }
>>
>>  function get_patch_id_by_series_id_and_sha() {
>
>> --
>> 2.41.0
>>


      reply	other threads:[~2024-01-19 14:51 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-12-12 16:08 Aaron Conole
2024-01-11 18:37 ` Aaron Conole
2024-01-11 19:01   ` Ilya Maximets
2024-01-12 17:53     ` Aaron Conole
2024-01-11 19:36   ` Michael Santana
2024-01-15 20:00 ` Michael Santana
2024-01-19 14:51   ` Aaron Conole [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=f7t7ck59y16.fsf@redhat.com \
    --to=aconole@redhat.com \
    --cc=ci@dpdk.org \
    --cc=dceara@redhat.com \
    --cc=echaudro@redhat.com \
    --cc=i.maximets@ovn.org \
    --cc=msantana@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).