Tuning Cloud services parameters

This topic describes the tunable parameters for Cloud services and the commands to modify them.

Cloud services use some default configuration parameters. You can change the value of the parameters if the default settings do not suit your requirements.

The following table provides the list of configurable parameters and their description.

Table 1. Attributes and default values
Variable Name Default Value Minimum Value Maximum Value Description
connector.server.timeout 5000 (ms) 1000 (ms) 15000 (ms) This is the maximum amount of time the server takes to respond to the client request. If the request is not fulfilled, it closes the connection.
connector.server.backlog 0 0 100 The maximum queue length for incoming connection indications (a request to connect) is set to the backlog parameter. If a connection indication arrives when the queue is full, the connection is refused.
destroy.sql.batchsize 8196 8196 81960 Page size per delete local database objects operation.
destroy.cloud.batchsize 256 256 81960 Page size per cloud objects delete operation.
reconcile.sql.batchsize 8196 8196 81960 Reconcile processes files in batches. This parameter controls how many files are processed in a batch.
commands.reconcile.lockwait.timeout.sec 360(s) 60(s) 3600(s) Maximum time to acquire the lock on directory for the reconcile operation.
threads.gc.batchsize 4096 4096 40960 Page size of the Garbage Collector thread. We can increase this in case the memory usage is more.
migration.downgrade.lock.threshold.size.mb 64 (MB) 1 (MB) 64 (MB) Sets the size threshold on files for which the lock downgrade is completed. To save time, a lock downgrade is not completed on shorter files that can transfer quickly. For larger files, a lock downgrade is suggested because migration might take a long time.
cloud-retention-period-days 30 0 2147483647 Number of days for which the migrated data needs to be retained on the cloud after its file system object has been deleted or reversioned.
connector.server.migrate-threadpool-size 32 1 64 Thread pool size of the migration threads. User can do this by making sure the CPU resources (CPU speed and number of cores) of the Cloud service nodes match.
connector.server.recall-threadpool-size 32 1 64 Thread pool size of recall threads. User can increase the number of recall threads to improve the performance.
tracing.enable true true False Enables administrators to print trace messages of the internal components in a file. Controls the level of messages such as Info, Warning, or Error.
tracing.level ALL=4 See Table 2 See Table 2 Tracing level is to set non-default tracing levels for various Transparent Cloud Tiering internal components to generate more debug data if any problems occur.
audit.enable true true true Enables or disables auditing information.
threads.cut-slow.sleep.ms 6000000 (ms) 60000 (ms) 2^63 Sleep time between two slow cloud update thread runs.
threads.cut-slow.sizediff 268435456 (Bytes) 1048576 (Bytes) 2^63 Size threshold of Cloud Updater Slow Threads.
threads.cut-slow.timediff.ms 604800000 (ms) 86400000 (ms) 2^63 Time threshold of Cloud Updater Slow Threads to update the cloud database.
threads.cut-fast.sizediff 16777216 (Bytes) 1048576 (Bytes) 2^63 Size threshold of Cloud Updater Fast Threads.
threads.cut-fast.timediff-active.ms 1800000 (ms) 60000 (ms) 2^63 Active time of Cloud Updater Fast Threads.
Table 2. Supported Components
Variable Name Default Value
THRD Threading, Thread Pools
STCK The modular stack and how operations propagate up and down
BSCN operations related to the blob store connection
STAT Deals with States
MPAU Multi-part upload
CNCT Connectors
SLCE Slices
KMGR Key Management
ENCR Encryption
GCON Scale back end connector
ETAG Etag Based Integrity
MFST Manifest related Operations
PAYL Payload related Operations
ENVR Environment related operations
CDIR Cloud Directory Component Operations
RECN Reconcile
SCAN Scan operations
POLI GPFS Policy operations
AUTH Authentication
SQLL SQL operations
JRNL Journal operations
NOTF GPFS Event Notifications
MTRX Metrics Code
CDAM Core Data Model
GDAM GPFS Data Model
MTTV TCT TTV Validator
CONF TCT configuration Layer related operations.
SERV Serviceability
MMON Monitoring
To tune Cloud services parameters, issue a command according to this syntax:
mmcloudgateway config set --cloud-nodeclass CloudNodeClass  
                           Attributee=value[,Attributee=value...]
where,
  • <--cloud-nodeclass> is the node class configured for the Cloud services nodes.
  • <Attribute> is the value of the attribute that is provided.
For more information, refer to the following examples:
  • To set the value of the tconnector.server.timeout attribute to 10 seconds, issue this command:
    mmcloudgateway config set --cloud-nodeclass tct1 connector.server.timeout=10000
  • To reset the value of the tconnector.server.timeout attribute to the default value, issue this command:
    mmcloudgateway config set --cloud-nodeclass tct1 connector.server.timeout=DEFAULT
  • To set the tracing levels, issue this command:
    mmcloudgateway config set --cloud-nodeclass tct tracing.level=GCON=5:AUTH=4
  • To reset the value of the tracing components to the default value, issue this command:
    mmcloudgateway config set --cloud-nodeclass tct tracing.level=default
  • To reset the values of multiple attributes, issue this command:
    
mmcloudgateway config set --cloud-nodeclass tct tracing.level=CDIR=1:
JRNL=1:SQLL=1:RECN=1