|
|
|
|
Since many businesses require non-stop operation of their mission critical applications, ASC supports the following to help you protect the data in your storage network:
ASC's failover feature provides high availability for ASC operations by eliminating the down time that can occur should an ASC Server (software or hardware) fail.
Note: You will need to know the IP address(es) of the primary server (and the secondary server if you are configuring a mutual failover scheme). You will also need the health monitoring IP address(es). It is a good idea to gather this information and find available IP addresses before you begin the setup.
You will see a screen that shows you a status of options on your server.
Before a failover configuration can be created, the ASC system needs to know the ownership of each physical device for the selected server. Therefore, we recommend you allow the wizard to rescan the server's devices.
If you have recently used the Rescan option to rescan the selected server's physical adapters, you can skip the server scanning process.
If you are using ASC NAS and you get an error message that there is a conflict with user/group IDs on the secondary server, you should right-click on the Windows Client object (under NAS Clients) and select Properties to determine what range of user/group IDs is being used on the primary server. You should then go to the appropriate system file on the server that will become the secondary server and manually resolve the conflicting user/group IDs:
If there are conflicts, you will see a window with mismatched devices listed. For example, if you have a RAID array and one server sees all eight devices and the other server sees only four devices, you will see the devices listed here as mismatched.
You must resolve the mismatch before continuing. For example, the QLogic driver may not have loaded on one server and you will have to load it before going on.
If you fixed any mismatched devices in the last step, you will need to rescan before the wizard can continue.
If you are re-running the Failover wizard because you made a change to a physical device on one of the servers, you should rescan before continuing.
If you didn't have any conflicts and you have recently used the Rescan option to rescan the selected server's physical adapters, you can skip the scanning process.
Note: If this is the first time you are setting up a failover configuration, you will get a warning message if there are any logical resources (SAN, NAS, or Replica Resources), Snapshot Groups, or clients assigned to the secondary server. You will need to remove them and then restart the failover wizard.
Note: If you change the Server IP addresses while the Console is connected using those IP addresses, then the Failover wizard will not be able to successfully create the configuration.
If you uncheck the Include this Network Adapter for failover box, the wizard will display the next card it finds. You must choose at least one.
Note: In the event of a failover, you should use only those IP addresses that are configured as part of the failover configuration to connect to the server in the Console.
The health monitoring IP address remains with the server in the event of failure so that the server's health can be continually monitored.
We recommend that you use static IP addresses.
The proper adapter is usually selected for you, but you should confirm that the adapter shown is not the initiator on your secondary server that is connected to the storage array, and also that it is not the target adapter on your secondary server.
If you are configuring a mutual failover, you will need to set up the standby adapter for the secondary server as well.
Once your configuration is complete, each time you connect to either server in the Console, you will automatically be connected to the other as well.
You can see the current status of your failover configuration, including all settings, by checking the Failover Information tab for the server.
Failover events are also written to the primary server's Event Log, so you can check there for status and operational information, as well as any errors. You should be aware that when a failover occurs, the Console will show the failover partner's Event Log for the server that failed.
Mirroring provides high availability for ASC operations by minimizing the down time that can occur if a physical disk fails. The mirror can be defined with disks that are not necessarily identical to each other in terms of vendor, type, or even interface (SCSI, FC, iSCSI, Infiniband).
With mirroring, the primary disk is the disk that is used to read/write data for a SAN Client and the mirrored copy is a copy of the primary. Both disks are attached to a single ASC Server and are considered a mirrored pair. If the primary disk fails, the disks swap roles so that the mirrored copy becomes the primary disk.
ASC has two Mirroring options, Synchronous Mirroring and Asynchronous Mirroring.
ASC's Synchronous Mirroring option offers the ability to define a synchronous mirror for any ASC managed disk (virtualized or service-enabled).
In the Synchronous Mirroring design, each time data is written to a designated disk, the same data is simultaneously written to another disk. This disk maintains an exact copy of the primary disk. In the event that the primary disk is unable to read/write data when requested to by a SAN Client, ASC seamlessly swaps data functions to the mirrored copy disk.
The Synchronous Mirroring design has the added benefit of improving performance, because data is read from both mirrored disks.
ASC's Asynchronous Mirroring Option offers the ability to define a near real-time mirror for any ASC-managed disk (virtual or service-enabled) over long distances between data centers.
When you configure an asynchronous mirror, you create a dedicated cache resource and associate it to an ASC-managed disk. Once the mirror is created, ASC synchronizes the primary and secondary disks. This process does not involve the application server. After the synchronization is complete, all write-requests from the associated application server are sequentially delivered to the dedicated cache resource. This data is then committed to both the primary and its mirror as a separate background process. For added protection, the cache resource can also be mirrored.
You can enable mirroring for a single SAN or NAS Resource or you can use the batch feature to enable mirroring for multiple SAN or NAS Resources. You can also enable mirroring for an existing snapshot resource, cache resource, or replica resource.
For multiple SAN or NAS Resources, right-click on the SAN Resources or NAS Resources object and select Mirror --> Add.
For an existing snapshot resource or cache resource, right-click on the SAN or NAS Resource and select Snapshot Resource or Cache Resource --> Mirror --> Add.
Custom lets you select which physical device(s) to use and lets you designate how much space to allocate from each.
Express automatically creates the Mirrored Copy using the criteria you select:
- Select different drive - ASC will look for space on another hard disk.
- Select drives from different adapter/channel - ASC will look for space on another hard disk only if it is on a separate adapter/channel.
- Select any available drive - ASC will look for space on any disk, including the original. This option is useful if you have mapped a device (such as a RAID device) that looks to ASC like a single physical device.
You need a cache resource if you will be using Asynchronous Mirroring. If you do not have one, you can create it through the wizard.
The cache resource wizard will be launched automatically when you configure Asynchronous Mirroring but you do not have a cache resource. You can also create a cache resource by right-clicking on a SAN or NAS Resource and selecting Cache Resource --> Enable. For multiple SAN or NAS Resources, right-click on the SAN Resources or NAS Resources object and select Cache Resource --> Add.
Custom lets you select which physical device(s) to use and lets you designate how much space to allocate from each.
Express automatically creates the cache resource using the criteria you select:
- Select different drive - ASC will look for space on another hard disk.
- Select drives from different adapter/channel - ASC will look for space on another hard disk only if it is on a separate adapter/channel.
- Select any available drive - ASC will look for space on any disk, including the original. This option is useful if you have mapped a device (such as a RAID device) that looks to ASC like a single physical device.
You can now mirror your cache resource by highlighting the SAN or NAS Resource and selecting Cache Resource --> Mirror --> Add.
You can see the current status of your mirroring configuration by checking the General tab for a mirrored SAN Resource.
The concept of performing a snapshot is similar to taking a picture. When we take a photograph, we are capturing a moment in time and transferring this moment in time to a photographic medium, even while changes are occurring to the object we focused our picture on. Similarly, a snapshot of an entire device allows us to capture data at any given moment in time and move it to either tape or another storage medium, while allowing data to be written to the device.
The basic function of the snapshot engine is to allow point-in-time, "frozen" images to be created of data volumes (virtual drives) using minimal storage space. The snapshot initially uses no disk space. As new data is written to the source volume, the old data blocks are moved to a temporary snapshot storage area. By combining the snapshot storage with the source volume, the data can be recreated exactly at it appeared at the time the snapshot was taken. For added protection, a snapshot resource can also be mirrored through ASC.
A trigger is an event that tells ASC when it is time to perform a snapshot of a virtual device. ASC's Replication, TimeMark, Snapshot Copy, and Zero-Impact Backup options all trigger snapshots.
Each SAN and NAS Resource can have one Snapshot Resource. All of the ASC options that use Snapshot (Replication, TimeMark, Snapshot Copy, and Zero-Impact backup) share the same Snapshot Resource for the SAN/NAS Resource.
Each snapshot initially uses no disk space. As new data is written to the source volume, the old data blocks are moved to the Snapshot Resource. Therefore, it is not necessary to have 100% of the size of the SAN/NAS Resource reserved as a Snapshot Resource. The typical amount is around 1% to 5%, however, this depends on the amount of activity (number of new writes) to the disk. You can set the Snapshot Resource to expand automatically, as needed.
If you do not create a Snapshot Resource for your SAN/NAS Resource, when you configure Replication, TimeMark, Snapshot Copy, or backup, the Snapshot wizard will launch first, allowing you to create it.
You can create a snapshot resource for a single SAN or NAS Resource or you can use the batch feature to create snapshot resources for multiple SAN or NAS Resources:
For multiple SAN or NAS Resources, right-click on the SAN Resources or NAS Resources object and select Snapshot Resource --> Create.
Custom lets you select which physical device(s) to use and lets you designate how much space to allocate from each.
Express lets you designate how much space to allocate and then automatically creates a Snapshot Resource using all available devices.
- Select different drive - ASC will look for space on another hard disk.
- Select drives from different adapter/channel - ASC will look for space on another hard disk only if it is on a separate adapter/channel.
- Select any available drive - ASC will look for space on any disk, including the original. This option is useful if you have mapped a device (such as a RAID device) that looks to ASC like a single physical device.
If you want ASC to automatically expand the Snapshot Resource when space is running low, select the threshold level and check the Automatically allocate more space for the Snapshot Resource option.
You must also determine the amount of space to be allocated for each expansion. You can set this to be a specific size (in MB) or a percentage of the size of the Snapshot Resource. There is no limit to the number of times the Snapshot Resource can be expanded.
Once the low space threshold is triggered, the system will attempt to expand the resource by allocating additional space. The time required to accomplish this may be in milliseconds or even seconds, depending on how busy the system is.
During this time, if the space actually runs out, the Snapshot Resource's TimeMark policy (set on the next screen) takes effect, either deleting old TimeMarks or stopping data writes.
To avoid this, we recommend that you allow enough time for expansion after the low space threshold is reached to prevent space from running out.
We recommend that your safety margin be at least five seconds. This means that from the time the low space threshold is reached, while data is being written to the drive at maximum throughput, it will take a minimum of five seconds to fill up the rest of the drive. Therefore, if the maximum throughput is 50 MB/s, the threshold should be set for when the space is below 250 MB. Of course if the throughput is lower, the allowance can be lower accordingly.
This will only occur if you have reached the maximum allowable size for your Snapshot Resource or if you have chosen not to expand it.
If you select Stop writing data, the system will prevent any new writes from getting to the disk once the Snapshot Resource runs out of space and it cannot allocate any more. As a result, the Client can experience write errors. (The specific behavior is operating system and file system dependent. It may range from a simple error message to the client completely locking up.) If you do not want that to happen, you should select Delete the earlier snapshot automatically instead.
You will now see a new Snapshot tab for this SAN/NAS Resource.
You can see how much of your Snapshot Resource is currently being used and your expansion methods by checking the Snapshot tab for a SAN or NAS Resource.
Because Snapshot Resources record block-level changes, not file-level, you may not see the Usage Percentage decrease when you delete files. This is because deleted files really still exist on the disk.
You must have a Snapshot Resource for Snapshot Copy. If you do not have one, you will be prompted to create one.
Custom lets you select which physical device(s) to use and lets you designate how much space to allocate from each.
Express automatically creates the target for you from available hard disk segments.
Select Existing lets you select an existing resource. There are several restrictions as to what you can select:
- The target must be the same type (SAN or NAS) as the source.
- The target must be the same size as the source.
- The target cannot have any Clients assigned or attached.
Note: All data on the target will be overwritten.
You can see the current status of your Snapshot Copy by checking the General tab of both the virtual drive you copied from or copied to.
Snapshot Copy events are also written to the server's Event Log, so you can check there for status information, as well as any errors.
Building on ASC's snapshot engine technology, which ensures point-in-time consistency for data recovery purposes, the Snapshot Group feature allows virtual drives and service enabled drives to be grouped for snapshot synchronization purposes. Snapshots for all resources in a group are taken at the same time whenever a snapshot is triggered. This feature, working in conjunction with the database-aware Snapshot Agents, ensures transactional integrity for database files that reside on multiple disks.
When you create a snapshot group, you can configure TimeMark, Backup, and Replication for the entire group. All members of the snapshot group will get configured the same way.
Depending upon which options you enable, the subsequent screens will let you set group policies for those options. Refer to the appropriate section(s) (Replication, Zero-Impact backup, or TimeMark) for details on configuration.
If you select Replication, you will have to select a target server. This target server will get a snapshot group as well. When it comes to adding resources to a group configured for replication, you can add any resource that is already configured for replication on the target server or any resource that does not have replication configured at all. You cannot add a resource if it is configured for replication to a different server.
Each snapshot group can be comprised of multiple SAN or NAS Resources. Each resource can only join one group and you cannot have both types of resources in the same group.
There are several ways to add resources to a snapshot group. After you create a snapshot group, you will be prompted to add resources. At any time afterwards, you can:
Replication is the process by which a SAN or NAS resource maintains a remote copy of itself. At prescribed intervals, a snapshot is taken of the resource (known as the Primary disk) and the changed data is transmitted to the replica disk so that they are synchronized. The replica disk is usually located at a remote location. Under normal operation, clients do not have access to the replica disk.
If a disaster occurs and the replica is needed, the administrator can promote the replica to become the primary disk so that clients can access it. Replica disks can be configured for ASC's storage services, including TimeMark, which can be useful for viewing the contents of the disk.
ASC has two Replication options, IP Replication and Fibre Channel Replication.
IP Replication allows fast, data synchronization of storage volumes from one ASC Server to another over the IP network.
With IP Replication, the replica disk is located on a separate ASC Server, called the target server.
Fibre Channel Replication allows fast, data synchronization of storage volumes within one ASC Server. It can be used within metropolitan area Fibre Channel SANs, or can be used with IP-based Fibre Channel extenders.
With Fibre Channel Replication, the replica disk is connected to the ASC Server via a gateway using edge routers or protocol converters. Because there is only one ASC Server, the primary and target servers are the same server.
You can enable replication for a single SAN or NAS Resource or you can use the batch feature to enable replication for multiple SAN or NAS Resources.
You need Snapshot Resources for the primary and replica disks. If you do not have them, you can create them through the wizard.
For multiple SAN or NAS Resource, right-click on the SAN Resources or NAS Resources object and select Replication --> Add.
Each primary disk can only have one replica disk.
Note: If you are replicating a NAS Resource, you must have NAS running on the target server. It will also need to be running any time future replication occurs or if you promote or remove a replica disk.
For Fibre Channel Replication, select the Local Server.
For IP Replication, select any server but the Local Server.
If the server you want does not appear on the list, click the Add button.
You must select at least one policy, but you can have multiple.
If you enter a watermark value, when the value is reached, a snapshot will be taken and replication of that data will begin. If additional data (more than the watermark value) is written to the disk after the snapshot, that data will not be replicated until the next replication. If a replication that was triggered by a watermark fails, the replication will be re-started based on the retry value you enter, assuming the system detects any write activity to the primary disk at that time. Future watermark-triggered replications will not start until after another successful replication occurs.
If you are using ASC's I/O Accelerator, replication is triggered when the watermark value of data is moved from the cache resource to the disk.
The Start replication everyday at (hh:mm) field uses military time (00:00 - 23:59), where 00:00 represents midnight.
If you enter a value in the Schedule the replication every field, the replication will begin n hours/minutes from the time the replica disk is created.
Future replications will take place every n hours/minutes from the start of the previous scheduled replication.
If a replication is already occurring when the next time interval is reached, the new replication request will be ignored.
Note: if you are using ASC's Snapshot Agent for Microsoft Exchange 5.5, the time between each replication should be longer than the time it takes to stop and then re-start the database.
Custom lets you select which physical device(s) to use and lets you designate how much space to allocate from each.
Express automatically creates the replica for you from available hard disk segments.
Select Existing lets you select an existing resource. There are several restrictions as to what you can select:
- The target must be the same type (SAN or NAS) as the primary.
- The target must be the same size as the primary.
- The target can have Clients assigned to it but they cannot be connected during the replication configuration. Since ASC cannot detect users connected to NAS resources, you must make sure NAS users are not connected during the replication configuration. Otherwise, once the configuration is created, they will be disconnected.
Note: All data on the target will be overwritten.
If there is no Snapshot Resource on the target, the wizard will ask you to create one. Because old data blocks are moved to the Snapshot Resource as new data is written to the replica, the Snapshot Resource should be large enough to handle the amount of changed data that will be replicated. Since it is not always possible to know how much changed data will be replicated, it is a good idea for you to enable expansion on the target server's Snapshot Resource.
For added protection, you can now mirror your replica resource by highlighting the replica resource and selecting Mirror --> Add.
Note: Once you create your replication configuration, you should not change the hostname of the source (primary) server. If you do, you will need to recreate your replication configuration.
There are several ways to check replication status, from the Replica Resource on the target server, the Event Log, or from the Replication Status Report.
You can see the current status of your replication by checking the General tab for a Replica Resource on the target server.
If the Replication Status shows New, Unreplicated, the configuration has been created but no replication has taken place yet.
While this tab shows the status of your replication, you can see the status and policies set for replication by checking the Replication tab of the logical resource on the primary server.
Replication events are also written to the primary server's Event Log, so you can check there for status and operational information, as well as any errors.
The Replication Status Report can be run from the Reports object. It provides a centralized view for displaying real-time replication status for all drives enabled for replication. It can be generated for an individual drive, multiple drives, source server or target server, for any range of dates. This report is useful for administrators managing multiple servers that either replicate data or are the recipients of replicated data.
ASC's TimeMark option guards against soft errors, non-catastrophic data loss, including the accidental deletion of files and software/virus issues leading to data corruption. TimeMark protects where high availability configurations cannot, since in creating a redundant set of data, high availability configurations also create a duplicate set of soft errors by default. TimeMark protects data from your slip-ups, from the butter fingers of employees, unforeseen glitches during backup, and from the malicious intent of viruses.
The TimeMark option also provides an "undo button" for data processing. Traditionally, when an administrator performed operations on a data set, a full backup was required before each "dangerous" step, as a safety net. If the step resulted in undesirable effects, the administrator needed to restore the data set and start the process all over again. With ASC's TimeMark option, you can easily rollback (restore) a drive to its original state.
ASC's TimeView feature is an extension of the TimeMark option and allows you to mount a TimeMark as a virtual drive. Deleted files can be retrieved from the drive or the drive can be assigned to multiple application servers for concurrent, independent processing, all while the original data set is still actively being accessed/updated by the primary application server. This is useful for "what if" scenarios, such as testing a new payroll application on your actual, but not live, data.
Using TimeMark's Copy option, you can create a real, permanent, independent virtual drive from any TimeMark.
TimeMarks are point-in-time images of any SAN or NAS virtual drive. Using ASC's Snapshot technology, TimeMarks track multiple virtual images of the same disk marked by "time". If you need to retrieve a deleted file or "undo" data corruption, you can recreate/restore the disk instantly based on any of the existing TimeMarks.
With TimeMark's Advanced Backup feature, you can easily back up just those block-level changes represented by each TimeMark. This advanced backup feature enables standard third-party backup software to perform incremental and differential backups at the block-level, something that was previously impossible to do.
A Snapshot Resource is used to record block-level changes made to a virtual drive since the last TimeMark was triggered. Since TimeMark takes incremental snapshots of user data and only stores the changes to that data, the Snapshot Resource is only a small fraction of the original disk size.
You need a Snapshot Resource for the logical resource you are going to configure. If you do not have one, you will create it through the wizard.
ASC allows you to schedule periodic TimeMarks. For example, it is possible to create a TimeMark every hour, giving you the ability to restore the disk content back to a known good state with hourly granularity.
In order for a TimeMark to be created, you must select at least one policy. Otherwise, you will have enabled TimeMark, but not created any. You will then need to manually create them using TimeMark --> Create.
If you enter a value in the Schedule to create a TimeMark every field, a new TimeMark will be created in n hours/minutes. If you leave the Create the first TimeMark in field set to zero, the first TimeMark will be created as soon as the wizard completes.
The Create a TimeMark everyday at field uses military time (00:00 - 23:59), where 00:00 represents midnight.
The maximum number of TimeMarks that can be maintained is 256. The maximum does not include the snapshot images that are associated with TimeView resources. Once the maximum is reached, the earliest will be deleted.
You will now have a TimeMark tab for this SAN or NAS Resource.
You can see a list of TimeMarks for this virtual drive, along with your TimeMark policies, by checking the TimeMark tab.
Right-click on the virtual drive and select Refresh to update the information on this tab.
To see how much space TimeMark is using, check the Snapshot tab.
TimeView allows you to mount a TimeMark as a virtual drive. Use TimeView if you need to restore individual files from a drive but you do not want to rollback the entire drive to a previous point in time. Simply use TimeView to mount the TimeMark and then copy the files you need back to your original virtual drive.
TimeView also enables you to perform "what if" scenarios, such as testing a new payroll application on your actual, but not live, data. After mounting the TimeMark, the drive can be assigned to an application server for independent processing without affecting the original data set. A TimeView cannot be configured for any of ASC's storage services.
ASC's Zero-Impact Backup Enabler allows you to perform a local raw device tape backup/restore of your virtual drives.
A raw device backup is a low-level backup or full copy request for block information at the volume level. Linux's dd command generates a low-level request.
Using ASC's Zero-Impact Backup Enabler with raw device backup software eliminates the need for the application server to play any role in backup and restore operations. Application servers on the SAN benefit from performance increases and the elimination of overhead associated with backup/restore operations because the command and data paths are rendered exclusively local to the ASC Server. This results in the most optimal data transfer between the disks and the tape, and is the only way to achieve net transfer rates that are limited only by the disk's or tape's engine. The backup process automatically leverages ASC's snapshot engine to guarantee point-in-time consistency.
To ensure full transactional integrity, this feature integrates with ASC's Snapshot Agents and the Group Snapshot feature.
When used with ASC's TimeMark option, the Zero-Impact Backup Enabler can easily back up just those block-level changes represented by each TimeMark. This advanced backup feature enables standard third-party backup software to perform incremental and differential backups at the block-level, something that was previously impossible to do.
You must have a Snapshot Resource for each virtual device you want to back up. If you do not have one, you will be prompted to create one.
Use an existing TimeMark snapshot - (This option is only valid if you are using ASC's TimeMark option on this SAN/NAS resource.) If a TimeMark exists for this virtual device, that image will be used for the backup. It may or may not be a current image at the time backup is initiated. If a TimeMark does not exist, a snapshot will be taken.
Create a new snapshot - A new snapshot will be created for the backup, ensuring the backup will be made from the most current image.
Each time a backup is requested by a third-party backup application, ASC creates a backup session. Depending upon the snapshot criteria you set on the previous window, a snapshot may be taken at the start of the backup session. (If the resource is part of a snapshot group, snapshots for all resources in the group will be taken at the same time.) Subsequently, each raw device is opened for backup and then closed. Afterwards, your backup application may verify the backup by comparing the data on tape with that of the snapshot image that was created for this session. For this reason, it is important to maintain the backup session until the verification is complete. Because ASC cannot tell how long your backup application needs to rewind the tape and compare the data, you must select an option on this screen to tell ASC how long to maintain the session. Note that the session length only applies to backups (reading from a raw device), not restores (writing to a raw device), and the actual session will end within 60 seconds of the session length you specify.
Absolute session length - This option maintains the backup session for a set period of time from the start of the backup session. Use this option when you know approximately how long your backup should take. You can also use this option to limit the length of time that the backup can run.
Relative session length - This option maintains the backup session for a period of time after the backup has completed (the last raw device has been opened and closed). This is more flexible than the absolute session length since it may be difficult to estimate how long it takes to back up all devices. With relative time, you only need to estimate how long to wait after the last device is backed up. If there is a problem during the backup (for example, such as when a tape is full), and the backup cannot complete, the Inactivity timeout tells ASC how long to wait before ending the backup session.
 Acer, Inc. www.acer.com |
|
|
|
|