Problem Symptom
After a node is powered off, it is removed from the storage pool. However, the PlogClient configuration file, PolgServer configuration file, and EDS configuration file still exist on the node. After the node is powered on again, the pre-upgrade check fails. Error message: The EDS process on node xxx is abnormal.
Problem Diagnosis
Ensure that the node has been removed from the storage pool.
Run the following command on the removed node to check whether residual PlogClient configuration file /opt/fusionstorage/persistence_layer/vbs/conf/plogclient.conf exists on the node:
ll /opt/fusionstorage/persistence_layer/vbs/conf/plogclient.conf
If any query result is displayed, the PlogClient configuration file still exists.
Run the following command on the node to check whether the PlogSserver configuration file still exists:
ll /opt/fusionstorage/persistence_layer/osd/conf/osd_*_conf.cfg
If any query result is displayed, the PlogServer configuration file still exists.
Check whether the node belonged to the disk pool of the block service or the disk pool of the file service before the capacity reduction.
If the node belonged to the disk pool of the block service, run the following command to check whether the eds configuration file exists:
cat /opt/dsware/eds/eds/conf/ebs_server.conf 2>&1 | grep fsa_service_id
If the node belonged to the disk pool of the file service (NAS, HDFS), run the following command to check whether the eds-f configuration file exists:
cat /opt/dsware/eds/eds-f/conf/ebs_server.conf 2>&1 | grep fsa_service_id
If any query result is displayed, the eds-f configuration file still exists.
Causes
After the node is powered off, FSM cannot delete its configuration file. You need to manually delete the residual configuration file. If residual configuration files exist, the pre-upgrade check fails.
Solution
Run the following command to clear the residual PlogClient configuration file:
rm -rf /opt/fusionstorage/persistence_layer/vbs/conf/plogclient.conf
Clear the residual PlogClient information.
Log in to the active management node as user dsware and run the following command to clear the residual PlogServer configuration file:
sh /opt/dsware/client/bin/dswareTool.sh --op clearNode -type osd -ip ip -poolId id
Run the following command to clear the residual EDS information:
sh /opt/dsware/client/bin/dswareTool.sh --op clearNode -type eds -ip ip -poolId id
Check After Recovery
Check whether the /opt/fusionstorage/persistence_layer/osd/conf/osd_*_conf.cfg and /opt/fusionstorage/persistence_layer/vbs/conf/plogclient.conf files are deleted. Check whether the storage pool status and cluster status are normal and whether the pre-upgrade check is successful.
Suggestion and Summary
N/A
Applicable Versions
FusionStorage 8.0.1,OceanStor 100D 8.0.2,OceanStor 100D 8.0.3
