Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Add metric for showing the errant GTIDs in VTOrc #13281

Merged
merged 5 commits into from
Jun 15, 2023
Merged

Add metric for showing the errant GTIDs in VTOrc #13281

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
2443ec1
feat: add metric for errant gtid
GuptaManan100 Jun 9, 2023
08b3ebe
feat: update the errant gtid map always, so that once errant gtids ar…
GuptaManan100 Jun 9, 2023
088f559
feat: fix the test to match the new expectations
GuptaManan100 Jun 9, 2023
84098ee
feat: add mutex to protect the access to the map storing errant gtids
GuptaManan100 Jun 14, 2023
0fe73a4
feat: update the errant gtid map in a goroutine
GuptaManan100 Jun 14, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 20 additions & 0 deletions go/test/endtoend/vtorc/api/api_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,9 @@ limitations under the License.
package api

import (
"encoding/json"
"fmt"
"reflect"
"testing"
"time"

Expand Down Expand Up @@ -197,5 +199,23 @@ func TestAPIEndpoints(t *testing.T) {
require.NoError(t, err)
assert.Equal(t, 400, status, resp)
assert.Equal(t, "Filtering by shard without keyspace isn't supported\n", resp)

// Also verify that the metric for errant GTIDs is reporting the correct information.
_, resp, err = utils.MakeAPICall(t, vtorc, "/debug/vars")
require.NoError(t, err)
resultMap := make(map[string]any)
err = json.Unmarshal([]byte(resp), &resultMap)
require.NoError(t, err)
errantGTIDMap := reflect.ValueOf(resultMap["ErrantGtidMap"])
errantGtidTablets := errantGTIDMap.MapKeys()
require.Len(t, errantGtidTablets, 3)

errantGTIDinReplica := ""
for _, tabletKey := range errantGtidTablets {
if tabletKey.String() == replica.Alias {
errantGTIDinReplica = errantGTIDMap.MapIndex(tabletKey).Interface().(string)
}
}
require.NotEmpty(t, errantGTIDinReplica)
})
}
18 changes: 18 additions & 0 deletions go/vt/vtorc/inst/instance_dao.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,6 +31,7 @@ import (
"github.com/rcrowley/go-metrics"
"github.com/sjmudd/stopwatch"

"vitess.io/vitess/go/stats"
"vitess.io/vitess/go/vt/external/golib/sqlutils"

vitessmysql "vitess.io/vitess/go/mysql"
Expand All @@ -56,6 +57,12 @@ const (
var instanceReadChan = make(chan bool, backendDBConcurrency)
var instanceWriteChan = make(chan bool, backendDBConcurrency)

var (
// Mutex to protect the access of the following variable
errantGtidMapMu = sync.Mutex{}
errantGtidMap = make(map[string]string)
)

// Constant strings for Group Replication information
// See https://dev.mysql.com/doc/refman/8.0/en/replication-group-members-table.html for additional information.
const (
Expand Down Expand Up @@ -88,6 +95,11 @@ func init() {
_ = metrics.Register("instance.write", writeInstanceCounter)
_ = writeBufferLatency.AddMany([]string{"wait", "write"})
writeBufferLatency.Start("wait")
stats.NewStringMapFuncWithMultiLabels("ErrantGtidMap", "Metric to track the errant GTIDs detected by VTOrc", []string{"TabletAlias"}, "ErrantGtid", func() map[string]string {
errantGtidMapMu.Lock()
defer errantGtidMapMu.Unlock()
return errantGtidMap
})

go initializeInstanceDao()
}
Expand Down Expand Up @@ -429,6 +441,12 @@ Cleanup:
instance.GtidErrant, err = vitessmysql.Subtract(redactedExecutedGtidSet.String(), redactedPrimaryExecutedGtidSet.String())
}
}
// update the errant gtid map
go func() {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Curious why this is done in a goroutine rather than an inline sequential routine?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We run this code path of refreshing tablet data parallelly for all the tablets sometimes. In that case, I didn't want this function to be blocked. I think it is okay for them to spawn go-routines that sequentially update the map asynchornously from the function. This map is used only for the metric, so as long as the updates flow, its all okay (even if slightly delayed)

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I considered an alternate where instead of updating the map on this read, I would just read the information from the database on a time-tick. This approach has a slight advantage wherein it is faster in its updation.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That makes sense. Thank you!

errantGtidMapMu.Lock()
defer errantGtidMapMu.Unlock()
errantGtidMap[topoproto.TabletAliasString(tablet.Alias)] = instance.GtidErrant
}()
}

latency.Stop("instance")
Expand Down