The active (ISP or ISU) resource can have the following states.
| Primary | Mirror Primary | Resource is ‘up’ and its role is ‘Primary’. The resource is active and able to be read/written, replicating to targets. (res_state: ISP) |  |
| Primary | Mirror Primary | Resource is ‘up’ and its role is ‘Primary’. The LifeKeeper state is ISU or the local disk is not ‘UpToDate’. The local disk not ‘UpToDate” is usually the case where it is a ‘SyncTarget’ and an OSU system is the ‘SyncSource’. It is important to not stop or reboot the ‘SyncSource’ until synchronization is complete. (res_state: ISU) |  |
If the resource is ‘down’ on an ISP/ISU server then res_state will return ‘OSF’ with the standard icon and text.
The standby (OSU) resource can have the following states. These states are independent of the state of the resource on other nodes and most of the res_state values are made up of the specific status returned by DRBD (documented in DRBD User’s Guide, section 5.2.8 Disk states). The DRBD statuses are in bold:
| Primary | Mirror Primary | The resource is ‘up’ and its role is ‘Primary’. This is usually a transition state where the resource state has not been updated to ISP yet during restore. When seen other than during a restore this is an unexpected state that may indicate manual manipulation of the resource using DRBD command line interface. (res_state: Primary) |  |
| Resyncing | Mirror Resynchronizing | The resource is ‘up’ and its role is ‘Secondary’. The DRBD replication-state is ‘resyncing’, ‘SyncTarget’ or ‘SyncSource’. (res_state: Resync) |  |
| Paused | Mirror Paused | The resource is ‘up’ and its role is ‘primary’. There are no connections to other targets. A flag file differentiates this with a ‘Primary’ status above. (res_state: Paused) |  |
| Verifying | Online Mirror Verification | The resource is ‘up’ and its role is ‘secondary’. A verify operation is in-progress with no mismatched data found.The DRBD replication-state is ‘VerifyT’ or ‘VerifyS’. (res_state: Verify) |  |
| Verifying - Mismatch | Online Mirror Verification - Mismatch found | The resource is ‘up’ and its role is ‘secondary’. A verify operation is in-progress with mismatched data found. The DRBD replication-state is ‘VerifyT’ or ‘VerifyS’ where out-of-sync is not 0. (res_state: VerifyMismatch) |  |
| UpToDate - Out Of Sync | Disk UpToDate with Blocks Out Of Sync | The resource is ‘up’ and its role is ‘Secondary”. While the data on the disk is consistent and up-to-date it has blocks that are out-of-sync with other servers. This can occur due to unresolved split-brain or mismatched blocks found by verify. The resource will not have a connection to another server. (res_state: UpToDateOutOfSync) | |
| UpToDate - Connection Issue | Disk UpToDate with Connection Issue | The resource is ‘up’ and its role is ‘Secondary’. The resource is not properly connected to another server. (res_state: UpToDateWarn) | |
The standby (OSU) resource can have the following states. Some states exist when there is an ISP (Primary) resource in the cluster (Active), or when the resource is ‘up’ and there is no ISP (Primary) resource in the cluster (Up) or when the resource is ‘down’ on the standby server (Down). The table below will identify if the state is seen Active, Up, and/or Down.
| UpToDate | Disk UpToDate | The resource is ‘up’ and its role is ‘Secondary’. The data on the disk is consistent and up-to-date. This is the normal state. Active/Up (res_state: UpToDateActive, UpToDateUp) | |
| Diskless | No Local Block Device | The resource is ‘up’ and its role is ‘Secondary’. There is a problem connecting to the local block device. This may mean the disk was automatically detached after a lower-level I/O error or it has been manually detached using ‘drbdadm detach’. Active/Up (res_state: DisklessActive, DisklessUp) |  |
| Outdated | Mirror Out of Synchronization | The resource is ‘up’ and its role is “Secondary”. The data is consistent but is not up to date with the primary node. | |
| Inconsistent Inconsistent (down) | Mirror is Not Consistent | The resource may be ‘up’ or ‘down’. It can not be in the role of ‘Primary’. The data is not usable. Active/Up/Down (res_state: InconsistentActive, InconsistentUp, InconsistentDown | |
| Unknown | Mirror State Unknown | The resource is ‘down’ and the state of the data on the disk was not determined. Down (res_state: DUnknownDown) |  |
| Detaching | Detaching Block Device | The resource is ‘up’ but being taken ‘down’. This is a transient state while detaching and waiting for ongoing IO’s to complete. This will lead to the resource going ‘down’. Active/Up (res_state: DetachingActive, DetachingUp) | |
| Failed | Failed Block Device | The resource is ‘up’. This is a transient state following an IO failure report by the local block device. This will lead to the Diskless state. Active/Up (res_state: FailedActive, FailedUp) | |
| Negotiating | Attaching Block Device | An ‘up’ operation is in-progress. This is a transient state when attaching to an already connected DRBD device. Up (res_state: NegotiatingUp) |  |
| Attaching | Attaching Block Device | An ‘up’ operation is in-progress. This is a transient state while reading meta data. This will lead to a state showing the meta data status (Consistent, Inconsistent, etc). Up (res_state: AttachingUp) | |
| Consistent | Mirror Data is Consistent | The resource is ‘down’. The data on the disk is consistent and usable. When the resource is brought ‘up’ and connected to other servers it is decided if the data is UpToDate or Outdated. Down (res_state: ConsistentDown) | |
Post your comment on this topic.