openshift/installer
1
0
Fork 0
mirror of https://github.com/openshift/installer.git synced 2026-02-05 15:47:14 +01:00
Code Issues Releases Activity

OCPBUGS-4998: Add additional info in wait-for when status is pending-user-action

Browse Source 
When the cluster status is installing-pending-user-action the install
won't complete. Most likely this is due to an invalid boot disk. When
this status is detected also log the host's status_info for hosts that
have this status.
This commit is contained in:
Bob Fournier
2023-04-03 17:02:42 -04:00
parent 439f39c8ed
commit 21b5a9fbb2
2 changed files with 20 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
pkg/agent/cluster.go
View File

@@ -219,6 +219,16 @@ func (czero *Cluster) IsBootstrapComplete() (bool, bool, error) {
czero.PrintInstallStatus(clusterMetadata)
// If status indicates pending action, log host info to help pinpoint what is missing
if (*clusterMetadata.Status != czero.installHistory.RestAPIPreviousClusterStatus) &&
(*clusterMetadata.Status == models.ClusterStatusInstallingPendingUserAction) {
for _, host := range clusterMetadata.Hosts {
if *host.Status == models.ClusterStatusInstallingPendingUserAction {
logrus.Debugf("Host %s %s", host.RequestedHostname, *host.StatusInfo)
}
}
}
if *clusterMetadata.Status == models.ClusterStatusReady {
stuck, err := czero.IsClusterStuckInReady()
if err != nil {

15
pkg/agent/waitfor.go
View File

@@ -18,7 +18,8 @@ func WaitForBootstrapComplete(cluster *Cluster) error {
waitContext, cancel := context.WithTimeout(cluster.Ctx, timeout)
defer cancel()
var lastErr error
var lastErrOnExit error
var lastErrStr string
wait.Until(func() {
bootstrap, exitOnErr, err := cluster.IsBootstrapComplete()
if bootstrap && err == nil {
@@ -28,10 +29,14 @@ func WaitForBootstrapComplete(cluster *Cluster) error {
if err != nil {
if exitOnErr {
lastErr = err
lastErrOnExit = err
cancel()
} else {
logrus.Info(err)
if err.Error() != lastErrStr {
logrus.Info(err)
lastErrStr = err.Error()
}
}
}
@@ -47,10 +52,10 @@ func WaitForBootstrapComplete(cluster *Cluster) error {
waitErr := waitContext.Err()
if waitErr != nil {
if waitErr == context.Canceled && lastErr != nil {
return errors.Wrap(lastErr, "bootstrap process returned error")
if errors.Is(waitErr, context.Canceled) && lastErrOnExit != nil {
return errors.Wrap(lastErrOnExit, "bootstrap process returned error")
}
if waitErr == context.DeadlineExceeded {
if errors.Is(waitErr, context.DeadlineExceeded) {
return errors.Wrap(waitErr, "bootstrap process timed out")
}
}