Troubleshoot and Maintain Dell EMC PowerFlex v3.6.x This section contains troubleshooting events and suggested solutions connected to the PowerFlex system. Child Topics Troubleshooting PowerFlex This guide contains topics to help you troubleshoot your PowerFlex system. SDS may fail to restart after deleting all volumes in FG Storage Pool In cases where the maximum or near-maximum number of storage devices are configured to an SDS, the SDS may fail to restart after the admin user deletes all volumes from a Fine Granularity (FG) Storage Pool. The problem occurs in two scenarios: Add acceleration pools You can add Acceleration Pools to a Protection Domain, to accelerate Storage Pool performance. After SDC installation, the Storage VM does not start automatically In cases where the SVM does not start automatically after the SDC is installed, ensure that the SVMs are set to start automatically. Application server does not see a PowerFlex volume Perform the following checks if the application server does not see a PowerFlex volume. False positive alerts during VMware deployment False alerts indicating the deployment has failed may appear when deploying a five-node cluster using the PowerFlex plug-in deployment wizard. Enabling RFcache acceleration in a new node Enable RFcache acceleration in a new SDS node in a PowerFlex system. Acceleration device failure on the SDS Beginning with PowerFlex v3.6, bad blocks in an NVDIMM acceleration device on a Linux-based SDS node cause an acceleration device error on the SDS. The error impacts fine granularity acceleration on the SDS node one of two ways, depending on whether NVDIMM hardware awareness is configured on the node. S.M.A.R.T. hardware alerts are not displayed Hardware alerts are not displayed in the PowerFlex GUI. Network failure occurs during automated PowerFlex installation using the PowerFlex Installer When installing PowerFlex using the PowerFlex Installer, an error message occurs that indicates the network is inaccessible: Replacing a faulty accelerated storage device in PowerFlex Replace a faulty accelerated (cached) storage device in a PowerFlex system. An error occurs when unmapping a volume from an ESXi server In cases where an error occurs when unmapping volumes from an ESXi server and the Could not detach LUN message is displayed, use the following steps to correct the failure and unmap the volume: PowerFlex GUI or SCLI cannot connect to an MDM In cases where the PowerFlex GUI or SCLI cannot connect to an MDM try the following checks: PowerFlex Gateway failure due to port collision The PowerFlex Gateway fails to run, or it runs but does not listen to any communications. In some cases, gateway installation fails. Restart PowerFlex Gateway before configuring the Set Alert Password when the lockbox directory is deleted or corrupted If the lockbox directory is corrupted, lockbox-related features such as LDAP, SNMP, ESRS, and Alert Service will not work. If the lockbox directory was deleted, the /opt/emc/scaleio/gateway/conf/lck_xml does not exist. No alert appears in either case. Some PowerFlex Gateway log files fail to rotate Following an upgrade, the PowerFlex Gateway Apache Tomcat log rotations may not be set on a Linux-based node. To remedy the situation, manually configure the conf/logging.properties file to specify the maximum number of retention days for a log entry so that entries are rotated accordingly. SCLI add_sds command fails due to communication error or MDM going offline When you run the SCLI add_sds command, the command may fail either due to a communication error or the Master MDM going offline. The vSphere PowerFlex plug-in responds slowly To resolve a slow response by the PowerFlex plug-in, increase memory size and restart the vSphere web client service. "Next" button disabled in Flash plug-in during PowerFlex VMware installtion Due to a software issue, after entering valid values in the relevant fields of the Configure SVM page during PowerFlex VMware installation, the Next button remains grayed out. Using PowerFlex vSphere plugin following Adobe Flash Player end of life The PowerFlex vSphere plugin, used to manage PowerFlex hyperconverged systems, is based on Adobe Flash technology. This topic explains the different work-arounds for customers using the vSphere plugin following Adobe Flash Player end of life on December 31, 2020. Solving PowerFlex performance issues PowerFlex is designed to generate the best performance possible from any given system configuration, by using all possible nodes and distributing the data evenly among them. Mismatch in I/O counters Resolving mismatch in the I/O counters of the PowerFlex Dashboard and customer application. Deploying SVM on a node with a management IP address from a different subnet than the node with the SVM template Due to a VMware limitation, it is not possible to deploy an SVM on a node with a management IP address that is from a different subnet than the node that hosts the SVM template. Speed up rebuild and rebalance processes Set concurrent activity limits to speed up rebuild and rebalance processes. This is useful If you are planning maintenance, have a disaster recovery (DR) situation, or require faster rebuild/rebalance times. SSD devices are not recognized in the operating system In an ESXi Server with a RAID controller, SSD devices are not recognized correctly by the ESXi host and are not displayed as SSDs in the vSphere plug-in. RAID controller, battery, or disk device hardware errors In Dell PowerEdge nodes, use the iDRAC System Event Log to look for RAID controller, cache battery, or disk device hardware errors that may lead to data loss or data integrity issues. Issue with fix deployment configuration when SDC settings not changed Perform the following steps to solve error after using Fix deployment configuration without changing the SDC and/or the virtual IP settings. Delete a certificate from the truststore In cases where a certificate is no longer valid or needs to be replaced, use the following command to delete the certificate from the truststore. Change the LIA password The LIA password is assigned during PowerFlex deployment and should be stored in a safe location with other passwords for future reference. Use the following procedure to change the LIA password in cases where the password was forgotten or misplaced, or is not consistent across nodes. Workaround for insufficient storage space for logs on Linux machines running PowerFlex Workaround for insufficient storage space for logs on Linux machines running PowerFlex Workaround for error handling of adding SDS devices during installation of PowerFlex with DirectPass I/O mode on vCenter Server Workaround for error handling of adding SDS devices during installation of PowerFlex with DirectPass I/O mode on vCenter Server Deployment fails to add previously partitioned device PowerFlex deployment may fail at the Add Device phase if a storage device was used previously for a different purpose and was partitioned. To fix this problem, follow these steps: Unregister PowerFlex system from SRS You can unregister the PowerFlex system from the SRS gateway. Validate ConnectEMC Dial-Home Verify that ConnectEMC (SRS) is correctly configured and is communicating with Dell EMC Global Services by performing a dial-home test. Add NVDIMM capacity to existing Fine Granularity Storage Pools Add NVDIMM capacity to Fine Granularity Storage Pools, to prepare your system in advance for future upgrades of PowerFlex. If your system has an active alert for missing NVDIMM capacity, this procedure will clear the alert. You will not be able to upgrade your system to versions higher than v3.0.1 unless sufficient NVDIMMs are installed in the system. Remove SDS from replication Within a replication environment, an SDS cannot be removed if the SDR contains journal volumes. The RCG must be removed before removing the SDS from the system. Insufficient journal capacity in the Storage Pool It is essential that there be sufficient capacity in the Storage Pool for the replication journal. Recover replication after network outage When there is a network outage between two replication peer systems, the journal capacity can be consumed as the SDRs at the source store all of the I/Os in the journals.