Skip to content

HDDS-11380. Fixing the error message of node decommission to be more comprehensive #7155

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

Merged
adoroszlai merged 5 commits into from
Oct 29, 2024
Merged

HDDS-11380. Fixing the error message of node decommission to be more comprehensive #7155

Changes from 2 commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
b7adbd5
HDDS-11380. Fixing the error message of node decommission to be more …
VarshaRaviCV Sep 4, 2024
314313b
Fixing the error message based on review comments
VarshaRaviCV Sep 5, 2024
cad5208
Making changes based on review comments
VarshaRaviCV Sep 10, 2024
608b9eb
Change the unhealthy keyword to not in-service or not healthy
VarshaRaviCV Sep 10, 2024
cbd007f
Adding node count assertions for decomm and maintenance in existing t…
VarshaRaviCV Sep 17, 2024
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
18 changes: 14 additions & 4 deletions ...dds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/node/NodeDecommissionManager.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -391,17 +391,21 @@ private synchronized boolean checkIfDecommissionPossible(List<DatanodeDetails> d
int numDecom = dns.size();
List<DatanodeDetails> validDns = new ArrayList<>(dns);
int inServiceTotal = nodeManager.getNodeCount(NodeStatus.inServiceHealthy());
int unHealthyTotal = nodeManager.getNodeCount(NodeStatus.inServiceStale()) +
nodeManager.getNodeCount(NodeStatus.inServiceDead());
Comment thread
VarshaRaviCV marked this conversation as resolved.
Outdated
for (DatanodeDetails dn : dns) {
try {
NodeStatus nodeStatus = getNodeStatus(dn);
NodeOperationalState opState = nodeStatus.getOperationalState();
if (opState != NodeOperationalState.IN_SERVICE) {
numDecom--;
validDns.remove(dn);
LOG.warn("Cannot decommission " + dn.getHostName() + " because it is not IN-SERVICE");
Comment thread
VarshaRaviCV marked this conversation as resolved.
Outdated
}
} catch (NodeNotFoundException ex) {
numDecom--;
validDns.remove(dn);
LOG.warn("Cannot decommission " + dn.getHostName() + " because it is not found in SCM");
Comment thread
VarshaRaviCV marked this conversation as resolved.
Outdated
}
}

Expand Down Expand Up @@ -431,8 +435,9 @@ private synchronized boolean checkIfDecommissionPossible(List<DatanodeDetails> d
int reqNodes = cif.getReplicationConfig().getRequiredNodes();
if ((inServiceTotal - numDecom) < reqNodes) {
String errorMsg = "Insufficient nodes. Tried to decommission " + dns.size() +
" nodes of which " + numDecom + " nodes were valid. Cluster has " + inServiceTotal +
" IN-SERVICE nodes, " + reqNodes + " of which are required for minimum replication. ";
" nodes out of " + inServiceTotal + " IN-SERVICE HEALTHY and " + unHealthyTotal +
" UNHEALTHY nodes. Cannot decommission as a minimum of " + reqNodes +
Comment thread
VarshaRaviCV marked this conversation as resolved.
Outdated
" IN-SERVICE HEALTHY nodes are required to maintain replication after decommission. ";
LOG.info(errorMsg + "Failing due to datanode : {}, container : {}", dn, cid);
errors.add(new DatanodeAdminError("AllHosts", errorMsg));
return false;
Expand Down Expand Up @@ -545,17 +550,21 @@ private synchronized boolean checkIfMaintenancePossible(List<DatanodeDetails> dn
List<DatanodeDetails> validDns = dns.stream().collect(Collectors.toList());
Collections.copy(validDns, dns);
int inServiceTotal = nodeManager.getNodeCount(NodeStatus.inServiceHealthy());
int unHealthyTotal = nodeManager.getNodeCount(NodeStatus.inServiceStale()) +
nodeManager.getNodeCount(NodeStatus.inServiceDead());
for (DatanodeDetails dn : dns) {
try {
NodeStatus nodeStatus = getNodeStatus(dn);
NodeOperationalState opState = nodeStatus.getOperationalState();
if (opState != NodeOperationalState.IN_SERVICE) {
numMaintenance--;
validDns.remove(dn);
LOG.warn(dn.getHostName() + " cannot enter maintenance because it is not IN-SERVICE");
}
} catch (NodeNotFoundException ex) {
numMaintenance--;
validDns.remove(dn);
LOG.warn(dn.getHostName() + " cannot enter maintenance because it is not found in SCM");
}
}

Expand Down Expand Up @@ -595,8 +604,9 @@ private synchronized boolean checkIfMaintenancePossible(List<DatanodeDetails> dn
}
if ((inServiceTotal - numMaintenance) < minInService) {
String errorMsg = "Insufficient nodes. Tried to start maintenance for " + dns.size() +
" nodes of which " + numMaintenance + " nodes were valid. Cluster has " + inServiceTotal +
" IN-SERVICE nodes, " + minInService + " of which are required for minimum replication. ";
" nodes out of " + inServiceTotal + " IN-SERVICE HEALTHY and " + unHealthyTotal +
" UNHEALTHY nodes. Cannot enter maintenance mode as a minimum of " + minInService +

@Tejaskriya Tejaskriya Sep 6, 2024

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Same suggestions as above

" IN-SERVICE HEALTHY nodes are required to maintain replication after maintenance. ";
LOG.info(errorMsg + "Failing due to datanode : {}, container : {}", dn, cid);
errors.add(new DatanodeAdminError("AllHosts", errorMsg));
return false;
Expand Down