Question
Why is SCSI performance superior to that of IDE/ATA? Explain the reasons from an architectural perspective
Answer :
SCSI offers improved performance and expandability and compatibility options, making it suitable for high-end computers.
- Number of devices supported is 16
- SCSI architecture derives its base from the client-server relationship
- SCSI initiator, or a client, sends a request to a SCSI target, or a server.
- The target performs the tasks requested and sends the output to the initiator
- When a device is initialized, SCSI allows for automatic assignment of device IDs on the bus, which prevents two or more devices using the same SCSI IDs
Question
What is a difference between a Cluster and a geographically-dispersed Cluster from administrative perspective?
Answer :
Geographically dispersed clusters, also called stretched clusters or extended clusters, are clusters comprised of nodes that are placed in different physical sites. Geographically dispersed clusters are designed to provide failover in the event of a site loss due to power issues, natural disasters or other unforeseen events.
From administrative perspective the difference would come up due to the storage that will be used. It won't be a common storage available at the respective locations instead a replication between the two will have to be set up and managed accordingly. Managing failover will also be different than a normal
cluster.
Question
DAS provides an economically viable alternative to other storage networking Solutions. Justify this statement
Answer :
- Setup requires a relatively lower initial investment
- Setup is managed using host-based tools, such as the host OS, which makes storage management tasks easy for small and medium enterprises.
- Requires fewer management tasks, and less hardware and software elements to set up and operate.
Question
i.Write the type of networks in place of N1 and N2. Write the type of ports in place of P1 and P2.
ii. What is meant by FC network is lossless? How FC achieve this ? How can we achieve losslessness in FCoE?
Answer
(i)N1 – IP Network. N2 – FC SAN. P1- Native SCSI port. N2- FC port
(ii) An FC network is lossless, meaning that the protocol has a built-in mechanism that prevents frame drops caused by congestion. Fibre Channel manages congestion through link level, credit based flow control. With credit-based flow control, the receiver sends credits to the sender to indicate the availability of receive buffers; the sender waits for credits before transmitting messages to the receiver Busy receive port can send the control frame to the transmit port for pause in transmission. This is called PAUSE capability of ethernet. Using this FCoE supports losslessness which is required in FC transmission.
Question
We have 6 nodes running a cluster. If suddenly 5 nodes found that they can communicate with each other but they cannot communicate with one specific node.
i. What steps the cluster should take to prevent data corruption? What is this phenomena called?
ii. Now if it is found that cluster is split in two groups with 3 nodes in each group. Nodes in one group can communicate with each other but can not communicate with nodes of other group. What is this situation called? Explain in brief about the steps the cluster will take to resolve the problem.
Answer :
(i) The node will be forced to shut down through some managed Switch. This is called fencing.
(ii)This is called as split brain. To prevent data corruption cluster should shut down the group with lesser number of nodes. Since here the number of nodes are equal, cluster will take the help of quorum disk to decide the group of nodes to be shut down
Question
A host generates 8,000 I/Os at peak utilization with an average I/O size of 32 KB. The response time is currently measured at an average of 12 ms during peak utilizations. When synchronous replication is implemented with a Fiber Channel ink to a remote site, what is the response time experienced by the host if the network latency is 6 ms per I/O?
Answer :
Actual response time = 12+ (6*4) + (32*1024/8000) = 40.096
Where 12 ms = current response time
6 ms per I/O = latency
32*1024/8000 = data transfer time
Question
It is required to connect one FOCE SAN and one FC SAN to a rack mounted servers having 10Gbe CNAs. Suggest a plan of connection with a diagram showing necessary components.
Answer :
Question
Explain the action involved between the NDMP DMA control and NDMP Server during Recovery process in the given scenario.
Answer :
DMA creates a control connection to the secondary storage agent
Connect using TCP port 10,000
NDMP_CONNECT_OPEN (to negotiate version)
NDMP_CONNECT_CLIENT_AUTH (to authenticate DMA to Server)
DMA uses the tape library media changer to load the required tape
The SCSI service is invoked
NDMP_SCSI_OPEN
NDMP_SCSI_EXECUTE_CDB - to manipulate media changer
NDMP_SCSI_CLOSE
DMA prepares the tape service for a recovery operation
The tape service is invoked
NDMP_TAPE_OPEN
NDMP_TAPE_READ - to validate volume label
NDMP_TAPE_MTIO - to position tape to start of backup data
DMA prepares the mover for a recovery operation
The mover is invoked
NDMP_MOVER_SET_RECORD_SIZE
NDMP_MOVER_SET_WINDOW
DMA opens control connection to the primary storage agent
Connect using TCP port 10,000
NDMP_CONNECT_OPEN - to negotiate protocol version
NDMP_CONNECT_CLIENT_AUTH - to authenticate DMA to Server
DMA queries secondary storage agent for capabilities
NDMP_CONFIG_GET_CONNECTION_TYPE
DMA queries primary storage agent for capabilities
NDMP_CONFIG_GET_BUTYPE_INFO
NDMP_CONFIG_GET_CONNECTION_TYPE
DMA obtains the data server’s data connection address information
The Data service is invoked
NDMP_DATA_LISTEN
DMA creates a data connection connection between NDMP servers
NDMP_MOVER_CONNECT
DMA creates a data connection connection between NDMP servers
The mover connects to the specified IP address & TCP port
DMA instructs the data server to initiate the recovery operation
NDMP_DATA_START_RECOVER
DMA recovery request is processed
Data service determines the offset & length of the DMA specified recovery data
Data server requests the specified data stream be transferred
NDMP_NOTIFY_DATA_READ
DMA instructs the mover to transfer the specified recovery stream
NDMP_MOVER_READ
the mover interacts with the tape service to access the recovery stream
DMA instructs the mover to transfer the specified recovery stream
The mover begins sending recovery stream over data connection
NDMP Data & Tape services send periodic log messages to DMA
NDMP_LOG_MESSAGE
NDMP Tape service sends notification when DMA intervention is required
example: end of mover window or tape medium encountered
NDMP_NOTIFY_MOVER_PAUSED
DMA initiates tape swap possibly utilizing media changer support
NDMP_TAPE_MTIO - to rewind/unload tape
NDMP_SCSI_EXECUTE_CDB - to manipulate media changer
NDMP_TAPE_MTIO - to position new tape
NDMP_TAPE_READ - to validate new tape header
DMA prepares the mover to continue the recovery operation
NDMP_MOVER_SET_WINDOW
NDMP_MOVER_CONTINUE
Data server detects end of recovery operation
Why is SCSI performance superior to that of IDE/ATA? Explain the reasons from an architectural perspective
Answer :
SCSI offers improved performance and expandability and compatibility options, making it suitable for high-end computers.
- Number of devices supported is 16
- SCSI architecture derives its base from the client-server relationship
- SCSI initiator, or a client, sends a request to a SCSI target, or a server.
- The target performs the tasks requested and sends the output to the initiator
- When a device is initialized, SCSI allows for automatic assignment of device IDs on the bus, which prevents two or more devices using the same SCSI IDs
Question
What is a difference between a Cluster and a geographically-dispersed Cluster from administrative perspective?
Answer :
Geographically dispersed clusters, also called stretched clusters or extended clusters, are clusters comprised of nodes that are placed in different physical sites. Geographically dispersed clusters are designed to provide failover in the event of a site loss due to power issues, natural disasters or other unforeseen events.
From administrative perspective the difference would come up due to the storage that will be used. It won't be a common storage available at the respective locations instead a replication between the two will have to be set up and managed accordingly. Managing failover will also be different than a normal
cluster.
Question
DAS provides an economically viable alternative to other storage networking Solutions. Justify this statement
Answer :
- Setup requires a relatively lower initial investment
- Setup is managed using host-based tools, such as the host OS, which makes storage management tasks easy for small and medium enterprises.
- Requires fewer management tasks, and less hardware and software elements to set up and operate.
Question
i.Write the type of networks in place of N1 and N2. Write the type of ports in place of P1 and P2.
ii. What is meant by FC network is lossless? How FC achieve this ? How can we achieve losslessness in FCoE?
Answer
(i)N1 – IP Network. N2 – FC SAN. P1- Native SCSI port. N2- FC port
(ii) An FC network is lossless, meaning that the protocol has a built-in mechanism that prevents frame drops caused by congestion. Fibre Channel manages congestion through link level, credit based flow control. With credit-based flow control, the receiver sends credits to the sender to indicate the availability of receive buffers; the sender waits for credits before transmitting messages to the receiver Busy receive port can send the control frame to the transmit port for pause in transmission. This is called PAUSE capability of ethernet. Using this FCoE supports losslessness which is required in FC transmission.
Question
i.Write Server Configuration for the following:
Export src and ports to client01 and client02, but only, client01 has root privileges on it.
The client machines have root and can mount anywhere on /exports. Anyone in the world can mount /exports/obj read-only.
ii.How do you recover the data from backup in following scenarios :
(a) Full backup taken on Monday, Incremental backup taken on Tuesday, Wednesday and Thursday. You have to restore system on Friday.
(b) Full backup taken on Monday, Cumulative backup taken on Tuesday, Wednesday and Thursday. You have to restore system on Friday.
Answer
i)
# Export src and ports to client01 and client02, but only
# client01 has root privileges on it
/usr/src /usr/ports -maproot=root client01
/usr/src /usr/ports client02
# The client machines have root and can mount anywhere
# on /exports. Anyone in the world can mount /exports/obj read-only
/exports -alldirs -maproot=root client01 client02
/exports/obj –ro
ii)
First restore Mondays full backup.
(a) Then restore backup of Tuesday, wednesday and Thursday.
(b) After restoration of Monday’s, restore Thursday’s backup
Question
A host generates 8,000 I/Os at peak utilization with an average I/O size of 32 KB. The response time is currently measured at an average of 12 ms during peak utilizations. When synchronous replication is implemented with a Fiber Channel ink to a remote site, what is the response time experienced by the host if the network latency is 6 ms per I/O?
Answer
Actual response time = 12+ (6*4) + (32*1024/8000) = 40.096
Where 12 ms = current response time
6 ms per I/O = latency
32*1024/8000 = data transfer time
We have 6 nodes running a cluster. If suddenly 5 nodes found that they can communicate with each other but they cannot communicate with one specific node.
i. What steps the cluster should take to prevent data corruption? What is this phenomena called?
ii. Now if it is found that cluster is split in two groups with 3 nodes in each group. Nodes in one group can communicate with each other but can not communicate with nodes of other group. What is this situation called? Explain in brief about the steps the cluster will take to resolve the problem.
Answer :
(i) The node will be forced to shut down through some managed Switch. This is called fencing.
(ii)This is called as split brain. To prevent data corruption cluster should shut down the group with lesser number of nodes. Since here the number of nodes are equal, cluster will take the help of quorum disk to decide the group of nodes to be shut down
Question
A host generates 8,000 I/Os at peak utilization with an average I/O size of 32 KB. The response time is currently measured at an average of 12 ms during peak utilizations. When synchronous replication is implemented with a Fiber Channel ink to a remote site, what is the response time experienced by the host if the network latency is 6 ms per I/O?
Answer :
Actual response time = 12+ (6*4) + (32*1024/8000) = 40.096
Where 12 ms = current response time
6 ms per I/O = latency
32*1024/8000 = data transfer time
Question
It is required to connect one FOCE SAN and one FC SAN to a rack mounted servers having 10Gbe CNAs. Suggest a plan of connection with a diagram showing necessary components.
Answer :
Question
Explain the action involved between the NDMP DMA control and NDMP Server during Recovery process in the given scenario.
DMA creates a control connection to the secondary storage agent
Connect using TCP port 10,000
NDMP_CONNECT_OPEN (to negotiate version)
NDMP_CONNECT_CLIENT_AUTH (to authenticate DMA to Server)
DMA uses the tape library media changer to load the required tape
The SCSI service is invoked
NDMP_SCSI_OPEN
NDMP_SCSI_EXECUTE_CDB - to manipulate media changer
NDMP_SCSI_CLOSE
DMA prepares the tape service for a recovery operation
The tape service is invoked
NDMP_TAPE_OPEN
NDMP_TAPE_READ - to validate volume label
NDMP_TAPE_MTIO - to position tape to start of backup data
DMA prepares the mover for a recovery operation
The mover is invoked
NDMP_MOVER_SET_RECORD_SIZE
NDMP_MOVER_SET_WINDOW
DMA opens control connection to the primary storage agent
Connect using TCP port 10,000
NDMP_CONNECT_OPEN - to negotiate protocol version
NDMP_CONNECT_CLIENT_AUTH - to authenticate DMA to Server
DMA queries secondary storage agent for capabilities
NDMP_CONFIG_GET_CONNECTION_TYPE
DMA queries primary storage agent for capabilities
NDMP_CONFIG_GET_BUTYPE_INFO
NDMP_CONFIG_GET_CONNECTION_TYPE
DMA obtains the data server’s data connection address information
The Data service is invoked
NDMP_DATA_LISTEN
DMA creates a data connection connection between NDMP servers
NDMP_MOVER_CONNECT
DMA creates a data connection connection between NDMP servers
The mover connects to the specified IP address & TCP port
DMA instructs the data server to initiate the recovery operation
NDMP_DATA_START_RECOVER
DMA recovery request is processed
Data service determines the offset & length of the DMA specified recovery data
Data server requests the specified data stream be transferred
NDMP_NOTIFY_DATA_READ
DMA instructs the mover to transfer the specified recovery stream
NDMP_MOVER_READ
the mover interacts with the tape service to access the recovery stream
DMA instructs the mover to transfer the specified recovery stream
The mover begins sending recovery stream over data connection
NDMP Data & Tape services send periodic log messages to DMA
NDMP_LOG_MESSAGE
NDMP Tape service sends notification when DMA intervention is required
example: end of mover window or tape medium encountered
NDMP_NOTIFY_MOVER_PAUSED
DMA initiates tape swap possibly utilizing media changer support
NDMP_TAPE_MTIO - to rewind/unload tape
NDMP_SCSI_EXECUTE_CDB - to manipulate media changer
NDMP_TAPE_MTIO - to position new tape
NDMP_TAPE_READ - to validate new tape header
DMA prepares the mover to continue the recovery operation
NDMP_MOVER_SET_WINDOW
NDMP_MOVER_CONTINUE
Data server detects end of recovery operation