Interview Questions.

Q1. What Is Luci And Ricci In Red Hat Cluster?

Luci is the server factor of the Conga administration utility

Conga is an included set of software components that offers centralized configuration and management of Red Hat clusters and garage

luci is a server that runs on one laptop and communicates with a couple of clusters and computer systems via ricci

ricci is the client factor of the Conga administration software

ricci is an agent that runs on every pc (either a cluster member or a standalone computer) controlled via Conga

This service desires to be walking on all of the client nodes of the cluster.

Q2. What Are The Various Types Of Fencing Supported By High Availability Add On?

Power fencing : fencing approach that makes use of a electricity controller to electricity off an inoperable node.

Storage fencing : A fencing approach that disables the Fibre Channel port that connects storage to an inoperable node.

Other fencing : Several other fencing techniques that disable I/O or strength of an inoperable node, consisting of IBM Bladecenters, PAP, DRAC/MC, HP ILO, IPMI, IBM RSA II, and others.

Q3. What Is The Order In Which You Will Start The Red Hat Cluster Services?

In Red Hat 4 :

carrier ccsd start

service cman start

service fenced begin

provider clvmd begin (If CLVM has been used to create clustered volumes)

carrier gfs begin

carrier rgmanager begin

In RedHat 5 :

service cman start

carrier clvmd start

carrier gfs start

carrier rgmanager start

In Red Hat 6 :

carrier cman start

service clvmd start

provider gfs2 begin

service rgmanager begin

Q4. How Many Nodes Are Supported In Red Hat 6 Cluster?

A cluster configured with qdiskd helps a most of sixteen nodes. The purpose for the restrict is due to scalability; growing the node depend will increase the amount of synchronous I/O competition at the shared quorum disk device.

Q5. What Is The Maximum File System Support Size For Gfs2?

GFS2 is primarily based on 64 bit structure, that could theoretically accommodate an eight EB file gadget. 

However, the modern-day supported most size of a GFS2 record machine for 64-bit hardware is a hundred TB. 

The present day supported most size of a GFS2 report gadget for 32-bit hardware for Red Hat Enterprise Linux Release 5.3 and later is 16 TB. 

NOTE: It is higher to have 10 1TB file structures than one 10TB report machine.

Q6. What Is Cman In Red Hat Cluster?

This is an abbreviation used for Cluster Manager. 

CMAN is a distributed cluster manager and runs in each cluster node. 

It is liable for tracking, heartbeat, quorum, vote casting and verbal exchange between cluster nodes.

CMAN maintains tune of cluster quorum by means of monitoring the count of cluster nodes.

Q7. What Is The Default Size Of Journals In Gfs?

When you run mkfs.Gfs2 with out the scale attribut for magazine to create a GFS2 partition, by means of default a 128MB length magazine is created that's sufficient for most of the programs

In case you intend on decreasing the dimensions of the magazine, it could significantly have an effect on the performance. Suppose you reduce the size of the magazine to 32MB it does no longer take plenty file gadget hobby to fill an 32MB magazine, and while the journal is complete, overall performance slows because GFS2 has to wait for writes to the garage.

Q8. What Is Dlm Lock Model?

DLM is a short abbreviation for Distributed Lock Manager.

A lock manager is a traffic cop who controls access to assets within the cluster, including get entry to to a GFS file system.

GFS2 uses locks from the lock manager to synchronize access to file device metadata (on shared garage)

CLVM makes use of locks from the lock manager to synchronize updates to LVM volumes and extent organizations (additionally on shared garage)

In addition, rgmanager uses DLM to synchronize carrier states.

Without a lock manager, there might be no control over get right of entry to to your shared storage, and the nodes within the cluster would corrupt every other's records.

Q9. What Are Tie-breakers In Red Hat Cluster?

Tie-breakers are additional heuristics that permit a cluster partition to decide whether or not or no longer it's miles quorate in the occasion of a good-break up - prior to fencing. 

With such a tie-breaker, nodes no longer handiest screen every other, however also an upstream router this is at the equal direction as cluster communications. If the 2 nodes lose touch with every other, the one that wins is the only which could nevertheless ping the upstream router.That is why, even when using tie-breakers, it's miles important to make sure that fencing is configured efficaciously.

CMAN has no internal tie-breakers for diverse reasons. However, tie-breakers may be implemented using the API.

Q10. What Is A High Performance Cluster?

High-overall performance clusters use cluster nodes to perform concurrent calculations. 

A excessive-performance cluster allows applications to work in parallel, therefore enhancing the performance of the applications. 

High performance clusters are also known as computational clusters or grid computing.

Q11. What Is High Availability Cluster?

High availability clusters provide exceedingly to be had offerings by using disposing of single factors of failure and by means of failing over services from one cluster node to some other in case a node turns into inoperative. 

Typically, offerings in a excessive availability cluster examine and write statistics (thru read-write set up report structures). 

A excessive availability cluster should keep information integrity as one cluster node takes over control of a carrier from some other cluster node. 

Node failures in a high availability cluster are not visible from customers out of doors the cluster. 

High availability clusters are every now and then called failover clusters.

Q12. What Are The Different Port No. Used In Red Hat Cluster?

IP Port no.

Protocol

Component

5404,5405

UDP

corosync/cman

11111

TCP

ricci

21064

TCP

dlm (Distributed Lock Manager)

16851

TCP

Modclustered

8084

TCP

luci

4196,4197

TCP

rgmanager

Q13. What Is The Minimum Size Of The Quorum Disk?

The minimal size of the block tool is 10 Megabytes.

Q14. What Is A Quorum Disk?

Quorum Disk is a disk-primarily based quorum daemon, qdiskd, that offers supplemental heuristics to determine node fitness.

With heuristics you could decide factors that are crucial to the operation of the node within the event of a community partition

For a three node cluster a quorum nation is present untill 2 of the three nodes are active i.E. More than half. But what if due to a few reasons the 2d node also stops speaking with the the 3rd node? In that case underneath a regular architecture the cluster could dissolve and stop operating. But for undertaking vital environments and such scenarios we use quorum disk in which an extra disk is configured which is established on all of the nodes with qdiskd service walking and a vote cost is assigned to it.

So think in above case I even have assigned 1 vote to qdisk so even after 2 nodes stops communicating with 3rd node, the cluster could have 2 votes (1 qdisk + 1 from third node) which is still more than half of vote matter for a 3 node cluster. Now both the inactive nodes could be fenced and your 3rd node could be nonetheless up and running being a part of the cluster.

Q15. What Is The Command Used To Relocate A Service To Another Node?

Clusvcadm -r service_name -m node_name

Q16. What Is Fencing In Red Hat Cluster?

Fencing is the disconnection of a node from the cluster's shared storage. 

Fencing cuts off I/O from shared garage, for that reason ensuring information integrity. 

The cluster infrastructure plays fencing thru the fence daemon, fenced.

When CMAN determines that a node has failed, it communicates to different cluster-infrastructure components that the node has failed. 

Fenced, whilst notified of the failure, fences the failed node. 

Q17. What Is Split-brain Condition In Red Hat Cluster?

We say a cluster has quorum if a majority of nodes are alive, communicating, and agree on the active cluster individuals. For instance, in a thirteen-node cluster, quorum is simplest reached if seven or extra nodes are speaking. If the 7th node dies, the cluster loses quorum and can not characteristic.

A cluster should preserve quorum to prevent break up-brain problems.

If quorum was now not enforced, quorum, a communication blunders on that same 13-node cluster can also motive a state of affairs where six nodes are running on the shared storage, at the same time as some other six nodes are also operating on it, independently. Because of the conversation error, the 2 partial-clusters would overwrite regions of the disk and corrupt the file system.

With quorum policies enforced, handiest one of the partial clusters can use the shared storage, for that reason protective data integrity.

Quorum would not save you cut up-brain situations, but it does determine who's dominant and allowed to function within the cluster.

Quorum may be determined by means of a mixture of speaking messages thru Ethernet and through a quorum disk.

Q18. What Is Storage Cluster?

Storage clusters provide a regular report device photo throughout servers in a cluster, allowing the servers to simultaneously study and write to a unmarried shared record gadget. 

A garage cluster simplifies storage administration with the aid of limiting the set up and patching of applications to at least one record system. 

The High Availability Add-On provides storage clustering together with Red Hat GFS2

Q19. What Are The Lock States In Red Hat Cluster?

A lock kingdom suggests the modern-day popularity of a lock request.

A lock is continually in one of three states:

Granted — The lock request succeeded and attained the requested mode.

Converting — A consumer attempted to exchange the lock mode and the new mode is incompatible with an existing lock.

Blocked — The request for a brand new lock couldn't be granted because conflicting locks exist.

A lock's state is decided by using its requested mode and the modes of the other locks on the identical resource.

Q20. What Is Rgmanager In Red Hat Cluster And Its Use?

This is a service termed as Resource Group Manager

RGManager manages and gives failover abilities for collections of cluster sources called offerings, useful resource organizations, or useful resource timber

it permits administrators to define, configure, and reveal cluster offerings. In the occasion of a node failure, rgmanager will relocate the clustered provider to another node with minimal carrier disruption

Q21. How Does Networkmanager Service Affects Red Hat Cluster?

The use of NetworkManager isn't always supported on cluster nodes. If you have mounted NetworkManager to your cluster nodes, you must either do away with it or disable it.

# provider NetworkManager prevent

# chkconfig NetworkManager off

The cman carrier will not start if NetworkManager is either walking or has been configured to run with the chkconfig command

Q22. What Are The Performance Enhancements In Gfs2 As Compared To Gfs?

Better performance for heavy usage in a single directory

Faster synchronous I/O operations

Faster cached reads (no locking overhead)

Faster direct I/O with preallocated files (supplied I/O length in all fairness large, consisting of 4M blocks)

Faster I/O operations in wellknown

Faster Execution of the df command, because of quicker statfs calls

Improved atime mode to reduce the number of write I/O operations generated by means of atime while in comparison with GFS

GFS2 supports the following features.

Prolonged report attributes (xattr)

the lsattr() and chattr() characteristic settings via popular ioctl() calls

nanosecond timestamps

GFS2 makes use of much less kernel memory.

GFS2 calls for no metadata generation numbers.

Allocating GFS2 metadata does not require reads. Copies of metadata blocks in multiple journals are managed via revoking blocks from the magazine earlier than lock release.

GFS2 consists of a far simpler log manager that knows not anything approximately unlinked inodes or quota changes.

The gfs2_grow and gfs2_jadd instructions use locking to prevent a couple of times jogging at the same time.

The ACL code has been simplified for calls like creat() and mkdir().

Unlinked inodes, quota modifications, and statfs modifications are recovered with out remounting the magazine.

Q23. What Is Load Balancing Cluster?

Load-balancing clusters dispatch network provider requests to more than one cluster nodes to balance the request load among the cluster nodes. 

Load balancing offers cost-powerful scalability because you could suit the range of nodes according to load necessities. If a node in a load-balancing cluster turns into inoperative, the burden-balancing software program detects the failure and redirects requests to different cluster nodes. 

Node failures in a load-balancing cluster are not visible from clients out of doors the cluster. 

Load balancing is to be had with the Load Balancer Add-On.

Q24. How Can You Define A Cluster And What Are Its Basic Types?

A cluster is two or more computer systems (called nodes or individuals) that paintings collectively to perform a venture.

There are four primary styles of clusters:

Storage

High availability

Load balancing

High performance

Q25. What Is The Order To Stop The Red Hat Cluster Services?

In Red Hat four :

provider rgmanager prevent

carrier gfs prevent

provider clvmd stop

provider fenced forestall

provider cman prevent

provider ccsd forestall

In Red Hat five

carrier rgmanager prevent

provider gfs prevent

service clvmd forestall

provider cman prevent

In Red Hat 6 :

carrier rgmanager forestall

provider gfs2 stop

carrier clvmd forestall

service cman prevent

Q26. What Is The Journaling Filesystem?

A journaling filesystem is a filesystem that keeps a unique file called a magazine this is used to repair any inconsistencies that occur as the end result of an incorrect shutdown of a laptop.

In journaling report structures, every time GFS2 writes metadata, the metadata is devoted to the magazine before it is placed into place. 

This ensures that if the machine crashes or loses power, you may get better all of the metadata when the journal is automatically replayed at mount time.

GFS2 requires one magazine for every node within the cluster that needs to mount the file machine. For example, when you have a sixteen-node cluster however want to mount simplest the file device from  nodes, you want best  journals. If you want to mount from a third node, you could always upload a magazine with the gfs2_jadd command.