hoernschen/dendrite
1
0
Fork 0
mirror of https://github.com/hoernschen/dendrite.git synced 2025-07-29 12:42:46 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Handle errors differently in the DeviceListUpdater (#2695)

Browse source 
`If a device list update goes missing, the server resyncs on the next
one` was failing because a previous test would receive a `waitTime` of
1h, resulting in the test timing out.
This now tries to handle the returned errors differently, e.g. by using
the default `waitTime` of 2s. Also doesn't try further users in the
list, if one of the errors would cause a longer `waitTime`.
This commit is contained in:
Till 2022-09-07 11:44:27 +02:00 • committed by GitHub
parent 847032df36
commit 440eb0f3a2
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
4 changed files with 44 additions and 19 deletions
Download patch file Download diff file Expand all files Collapse all files

48
keyserver/internal/device_list_update.go
View file

@ -19,9 +19,11 @@ import (
"encoding/json"
"fmt"
"hash/fnv"
"net"
"sync"
"time"
"github.com/matrix-org/gomatrix"
"github.com/matrix-org/gomatrixserverlib"
"github.com/matrix-org/util"
"github.com/prometheus/client_golang/prometheus"
@ -388,6 +390,8 @@ func (u *DeviceListUpdater) processServer(serverName gomatrixserverlib.ServerNam
return waitTime, true
}
failCount := 0
userLoop:
for _, userID := range userIDs {
if ctx.Err() != nil {
// we've timed out, give up and go to the back of the queue to let another server be processed.
@ -397,19 +401,35 @@ func (u *DeviceListUpdater) processServer(serverName gomatrixserverlib.ServerNam
res, err := u.fedClient.GetUserDevices(ctx, serverName, userID)
if err != nil {
failCount += 1
fcerr, ok := err.(*fedsenderapi.FederationClientError)
if ok {
if fcerr.RetryAfter > 0 {
waitTime = fcerr.RetryAfter
} else if fcerr.Blacklisted {
switch e := err.(type) {
case *fedsenderapi.FederationClientError:
if e.RetryAfter > 0 {
waitTime = e.RetryAfter
break userLoop
} else if e.Blacklisted {
waitTime = time.Hour * 8
} else {
// For all other errors (DNS resolution, network etc.) wait 1 hour.
waitTime = time.Hour
break userLoop
}
} else {
waitTime = time.Hour
logger.WithError(err).WithField("user_id", userID).Debug("GetUserDevices returned unknown error type")
case net.Error:
// Use the default waitTime, if it's a timeout.
// It probably doesn't make sense to try further users.
if !e.Timeout() {
waitTime = time.Minute * 10
logrus.WithError(e).Error("GetUserDevices returned net.Error")
break userLoop
}
case gomatrix.HTTPError:
// The remote server returned an error, give it some time to recover
if e.Code >= 500 {
waitTime = time.Minute * 10
logrus.WithError(e).Error("GetUserDevices returned gomatrix.HTTPError")
break userLoop
}
default:
// Something else failed
waitTime = time.Minute * 10
logger.WithError(err).WithField("user_id", userID).Debugf("GetUserDevices returned unknown error type: %T", err)
break userLoop
}
continue
}
@ -437,7 +457,11 @@ func (u *DeviceListUpdater) processServer(serverName gomatrixserverlib.ServerNam
}
}
if failCount > 0 {
logger.WithField("total", len(userIDs)).WithField("failed", failCount).WithField("wait", waitTime).Warn("Failed to query device keys for some users")
logger.WithFields(logrus.Fields{
"total": len(userIDs),
"failed": failCount,
"skipped": len(userIDs) - failCount,
}).Warn("Failed to query device keys for some users")
}
for _, userID := range userIDs {
// always clear the channel to unblock Update calls regardless of success/failure