PowerFlex: Slik tilbakestiller du oscillerende feil

Tilbakestille oscillerende feil

Oscillerende feil er kosmetiske og forårsaker ingen problemer. Når kunder utfører RCM-oppgraderinger, implementerer en ny tjeneste eller noe som sender verter opp og ned ofte nok, rapporterer PowerFlex dette som en oscillerende feil.

De ulike oscillerende feiltellerne kan tilbakestilles til null. Følgende kommando er nyttig når du har løst et problem og vil sikre at et varsel ikke lenger er aktivt i systemet. Vi kan tilbakestille tellere for hele systemet, per beskyttelsesdomene eller per lagringsutvalg.

• SSH til primær MDM for å kjøre tilbakestillingskommandoen nedenfor:

Syntaks:

scli --reset_oscillating_failure_counters (--failure_counter (((-- protection_domain_id | --protection_domain_name ) | --all_sds) | --all_sdc | (((----protection_domain_id | -- protection_domain_name) --storage_pool_name ) | --storage_pool_id ) | --all_devices) | --all_counters)

Parametere

--failure_counter <COUNTER>

Den oscillerende feiltelleren skal tilbakestilles til null; Tellere er gruppert som følger:

• Tellergruppe A (MDM- og SDS-relaterte tellere):

mdm_sds_network_disconnections

sds_sds_network_disconnections

sds_decoupled

sds_configuration_failures

sds_receive_buffer_allocation_failures

• Tellergruppe B (SDC-relaterte tellere):

sdc_mdm_network_disconnections

sdc_sds_network_disconnections

sdc_long_operations

sdc_memory_allocation_failures

sdc_socket_allocation_failures

• Tellergruppe C (SDS-enhetsrelaterte tellere):

sds_device_long_successful_ios

--protection_domain_id <ID>         Protection Domain ID

--protection_domain_name <NAME>   - Protection Domain name

--storage_pool_name <NAME>        - Storage Pool name

--storage_pool_id <ID>            - Storage Pool ID

--all_sds                         - Reset the counter parameters on all SDSs

--all_sdc                         - Reset the counter parameters on all SDCs

--all_devices                     - Reset the counter parameters on all SDS devices

--all_counters                    - Reset all counters to zero

Eksempel (alle SDC-, SDS-, MDM-tellere):

3.x 
scli --login --username admin --password <admin_password>


4.x 
scli --add_certificate --certificate_file /opt/emc/scaleio/mdm/cfg/mgmt_ca.pem      (Optional)
scli --login --username admin --password <admin_password> --management_system_ip <PFMP_UI_IP>

# scli --reset_oscillating_failure_counters --all_counters --all_sdc

# scli --reset_oscillating_failure_counters --all_counters --all_sds

# scli --reset_oscillating_failure_counters --all_counters --all_mdm

Eksempel (spesifikk teller):

# scli --log in --username admin –password <MDM Password>

# scli --reset_oscillating_failure_counters --failure_counter mdm_sds_network_disconnections --all_sds

Tilleggsinformasjon – spesifikke tellerdetaljer:

sds_configuration_failures              - measures the number of MDM's failures to configure an SDS when connecting to an SDS (failures in the reconfiguration phase)
sds_receive_buffer_allocation_failures  - measures the number of times SDS fails to allocate buffer for receiving messages

sdc_mdm_network_disconnections
sdc_sds_network_disconnections
sdc_long_operations                     - measures the number of SDC RPC operations which took more time than a predefined threshold
sdc_memory_allocation_failures          - measures the number of memory allocation failures in each SDC
sdc_socket_allocation_failures          - measures the number of socket allocation failures in each SDC

sds_device_long_successful_ios          - measures the number of successful IOs to an SDS device that took longer than a predefined threshold

Standard terskel er 250 millisekunder.