Abstract: A computer-implemented method according to one embodiment includes assigning a priority to virtual machines (VMs), and obtaining pathing information for the VMs. At least one of the VMs assigned a relatively higher priority is selected to perform a process over at least one of the VMs assigned a relatively lower priority. A computer program product according to another embodiment includes a computer readable storage medium having program instructions embodied therewith. The program instructions are readable and/or executable by a computer to cause the computer to perform the foregoing method. A system according to another embodiment includes a processor, and logic integrated with the processor, executable by the processor, or integrated with and executable by the processor. The logic is configured to perform the foregoing method.

Abstract: A method, computer system, and computer program product area provided. A computer transmits a query command to a storage descriptor area of a first disk. The first disk belongs to a dual-site data replication system. The dual-site data replication system provides active-active access to a volume of data stored in an active disk and replicated in a backup disk. The computer receives a response to the query command. The response indicates the active disk and the backup disk for the dual-site data replication system. The computer controls an additional copy of the volume of data at a further remote site based on the active disk.

Abstract: Examples of techniques for failure detection for central electronics complex (CEC) group management are described herein. An aspect includes issuing a first virtual input/output server (VIOS) probe to a hardware management console (HMC) of a central electronics complex (CEC) group. Another aspect includes receiving a first response packet that includes health data corresponding to a plurality of VIOSes. Another aspect includes determining, based on the first response packet, that cluster down is indicated on a first VIOS. Another aspect includes, based on determining that cluster down is indicated on the first VIOS, getting a VIOS state for the first VIOS from the HMC. Another aspect includes determining based on the VIOS state that the first VIOS is in a down state and determining that the first VIOS is unhealthy. Another aspect includes updating a health data entry corresponding to the first VIOS to indicate that the first VIOS is unhealthy.

Abstract: A write operation storing data in a first storage device is duplicated to a first replication file. A set of differences between a first version of the first replication file determined at a first time and a second version of the first replication file determined at a second time is determined, the set of differences comprising a set of results of duplicated write operations occurring between the first time and the second time. At a second file system, storage of the set of differences in a second storage device is caused, creating a duplicate in the second storage device of the data stored in the first storage device.

Abstract: In a system having at least two data storage and processing sites, each capable of alternatively serving as a primary site and a backup or target site, disaster recovery migration is optimized by cognitively analyzing at least one system parameter. Using machine learning, at least one pattern of that system related parameter is predicted, and planned or unplanned migration procedures are performed based on the predicted parameter patterns. The analyzed parameter may be data traffic at the sites, and the predicted data traffic pattern is used to assign primary and backup site status to those sites. The analyzed parameter may be the occurrence of events or transactions at the sites, and the predicted event or transaction patterns may be used to determine times of disaster recovery procedure processing so as to not interrupt a critical event or transaction.

Abstract: A system and method for automatic and dynamic virtual machine (VM) group membership adjustment to identify and maintain VM groupings for maintaining a data consistency and availability at a computer recovery site in response to a disaster event. Based on a hybrid approach of network and storage analytics using cognitive methods, the system automatically groups the VMs based on patterns of data exchanged between the VMs. The system can store the patterns found during the analysis, find the critical elements in those patterns (like VMs playing the key role of DB server), match the patterns and use them in case a detected failure to make the system functional again at the recovery site. The system stores the communications traffic patterns learned during the network analysis and uses analytics to form or change the VM grouping, and can perform the pattern matching to quickly readjust the group.

Abstract: A processor, based on detection of a configuration change to a central electronics complex (CEC) in a CEC group, automatically creates a monitoring network within the CEC. Automatically creating the monitoring network includes the processor issuing a request via a hardware management console (HMC) to the CEC to create a virtual switch implementing a virtual local area network (VLAN). The processor also issues a request via the HMC to the CEC to create, on a virtual input-output server (VIOS) hosted in the CEC, a virtual trunk adapter connected to the VLAN. In addition, the processor issues a request via the HMC to the CEC to create, on each of a plurality of logical partitions (LPARs) hosted in the CEC, a virtual network adapter connected to the VLAN. The processor can employ the network, for example, to monitor health of the LPARs and VIOS within the CEC via the HMC.

Abstract: Aspects include prefetching a plurality of high-level information, high-level metadata, low-level metadata, and low-level information including a plurality of components associated with a monitored entity for disaster recovery. A subsequent instance of the high-level information, the high-level metadata, and the low-level metadata is requested. The subsequent instance of the high-level information is fetched based on detecting a change in the high-level metadata. A subsequent instance of one or more of the components of the low-level information corresponding to one or more changes in the low-level metadata is fetched for updating a plurality of disaster recovery data of the monitored entity in a persistent database.

Abstract: A processor, based on detection of a configuration change to a central electronics complex (CEC) in a CEC group, automatically creates a monitoring network within the CEC. Automatically creating the monitoring network includes the processor issuing a request via a hardware management console (HMC) to the CEC to create a virtual switch implementing a virtual local area network (VLAN). The processor also issues a request via the HMC to the CEC to create, on a virtual input-output server (VIOS) hosted in the CEC, a virtual trunk adapter connected to the VLAN. In addition, the processor issues a request via the HMC to the CEC to create, on each of a plurality of logical partitions (LPARs) hosted in the CEC, a virtual network adapter connected to the VLAN. The processor can employ the network, for example, to monitor health of the LPARs and VIOS within the CEC via the HMC.

Abstract: In a system having at least two data storage and processing sites, each capable of alternatively serving as a primary site and a backup or target site, disaster recovery migration is optimized by cognitively analyzing at least one system parameter. Using machine learning, at least one pattern of that system related parameter is predicted, and planned or unplanned migration procedures are performed based on the predicted parameter patterns. The analyzed parameter may be data traffic at the sites, and the predicted data traffic pattern is used to assign primary and backup site status to those sites. The analyzed parameter may be the occurrence of events or transactions at the sites, and the predicted event or transaction patterns may be used to determine times of disaster recovery procedure processing so as to not interrupt a critical event or transaction.

Abstract: Examples of techniques for failure detection for central electronics complex (CEC) group management are described herein. An aspect includes issuing a first virtual input/output server (VIOS) probe to a hardware management console (HMC) of a central electronics complex (CEC) group. Another aspect includes receiving a first response packet that includes health data corresponding to a plurality of VIOSes. Another aspect includes determining, based on the first response packet, that cluster down is indicated on a first VIOS. Another aspect includes, based on determining that cluster down is indicated on the first VIOS, getting a VIOS state for the first VIOS from the HMC. Another aspect includes determining based on the VIOS state that the first VIOS is in a down state and determining that the first VIOS is unhealthy. Another aspect includes updating a health data entry corresponding to the first VIOS to indicate that the first VIOS is unhealthy.

Abstract: Examples of techniques for failure detection for central electronics complex (CEC) group management are described herein. An aspect includes issuing a first logical partition (LPAR) probe to a hardware management console (HMC) of a central electronics complex (CEC) group, wherein the CEC group comprises a plurality of LPARs. Another aspect includes receiving a first response packet from the HMC corresponding to the first LPAR probe, wherein the first response packet comprises health data corresponding to a first LPAR of the plurality of LPARs. Another aspect includes storing the health data corresponding to the first LPAR in a first health data entry corresponding to the first LPAR. Another aspect includes, for a second LPAR of the plurality of LPARs that was not included in the first response packet, updating a second health data entry corresponding to the second LPAR to indicate that the second LPAR is healthy.

Abstract: A system and method for automatic and dynamic virtual machine (VM) group membership adjustment to identify and maintain VM groupings for maintaining a data consistency and availability at a computer recovery site in response to a disaster event. Based on a hybrid approach of network and storage analytics using cognitive methods, the system automatically groups the VMs based on patterns of data exchanged between the VMs. The system can store the patterns found during the analysis, find the critical elements in those patterns (like VMs playing the key role of DB server), match the patterns and use them in case a detected failure to make the system functional again at the recovery site. The system stores the communications traffic patterns learned during the network analysis and uses analytics to form or change the VM grouping, and can perform the pattern matching to quickly readjust the group.

Abstract: Examples of techniques for failure detection for central electronics complex (CEC) group management are described herein. An aspect includes issuing a first logical partition (LPAR) probe to a hardware management console (HMC) of a central electronics complex (CEC) group, wherein the CEC group comprises a plurality of LPARs. Another aspect includes receiving a first response packet from the HMC corresponding to the first LPAR probe, wherein the first response packet comprises health data corresponding to a first LPAR of the plurality of LPARs. Another aspect includes storing the health data corresponding to the first LPAR in a first health data entry corresponding to the first LPAR. Another aspect includes, for a second LPAR of the plurality of LPARs that was not included in the first response packet, updating a second health data entry corresponding to the second LPAR to indicate that the second LPAR is healthy.

Abstract: Aspects include prefetching a plurality of high-level information, high-level metadata, low-level metadata, and low-level information including a plurality of components associated with a monitored entity for disaster recovery. A subsequent instance of the high-level information, the high-level metadata, and the low-level metadata is requested. The subsequent instance of the high-level information is fetched based on detecting a change in the high-level metadata. A subsequent instance of one or more of the components of the low-level information corresponding to one or more changes in the low-level metadata is fetched for updating a plurality of disaster recovery data of the monitored entity in a persistent database.

Abstract: In a system having at least two data storage and processing sites, each capable of alternatively serving as a primary site and a backup or target site, disaster recovery migration is optimized by cognitively analyzing at least one system parameter. Using machine learning, at least one pattern of that system related parameter is predicted, and planned or unplanned migration procedures are performed based on the predicted parameter patterns. The analyzed parameter may be data traffic at the sites, and the predicted data traffic pattern is used to assign primary and backup site status to those sites. The analyzed parameter may be the occurrence of events or transactions at the sites, and the predicted event or transaction patterns may be used to determine times of disaster recovery procedure processing so as to not interrupt a critical event or transaction.

Abstract: Two or more nodes in a cluster are determined. The cluster manages one or more resources found on a device. An indication from a first node of the two or more nodes is received. The indication includes a request for an IP lease for the first node. A parity bit setting for the first node is determined. Responsive to determining the parity bit setting of the first node being set to one, an IP lease to the first node is indicated. Responsive to determining the parity bit setting of the first node being set to zero, a denial of an IP lease to the first node is indicated.

Abstract: In a system having at least two data storage and processing sites, each capable of alternatively serving as a primary site and a backup or target site, disaster recovery migration is optimized by cognitively analyzing at least one system parameter. Using machine learning, at least one pattern of that system related parameter is predicted, and planned or unplanned migration procedures are performed based on the predicted parameter patterns. The analyzed parameter may be data traffic at the sites, and the predicted data traffic pattern is used to assign primary and backup site status to those sites. The analyzed parameter may be the occurrence of events or transactions at the sites, and the predicted event or transaction patterns may be used to determine times of disaster recovery procedure processing so as to not interrupt a critical event or transaction.

Abstract: Two or more nodes in a cluster are determined. The cluster manages one or more resources found on a device. An indication from a first node of the two or more nodes is received. The indication includes a request for an IP lease for the first node. A parity bit setting for the first node is determined. Responsive to determining the parity bit setting of the first node being set to one, an IP lease to the first node is indicated. Responsive to determining the parity bit setting of the first node being set to zero, a denial of an IP lease to the first node is indicated.

Abstract: Responsive to a cluster manager for a particular node from among multiple nodes allocating at least one leased resource for a resource group for an application workload on the particular node, on fallover of the resource group from another node to the particular node, setting a timer thread, by the cluster manager for the particular node, to track an amount of time remaining for an initial lease period of the at least one leased resource.