HP Integrity NonStop H-Series Guia do Utilizador descarregar pdf (Página 102)

HP NonStop Shadowbase Command Definitions

Collector Command Descriptions

HP NonStop Shadowbase Command Reference Manual—785421-002

LTS LAG (CUR) shows what the lag time is from the Collector’s current audit

trail read position (where the Collector is currently reading from) to the

current wall clock time (this is the current replication latency). This lag

time show how far behind the Collector is in audit trail processing. Note

that even when the Collector is caught up, the LTS lag may show that the

Collector is behind by up to ADTSAMPLEDELAY seconds (see below).

To see the Consumer LAG details, enter STATS COLL <collname>,

DETAIL. In addition to the lag times, the number of active transactions

being tracked by that Consumer, the number of messages queued to be

sent (i.e., not yet sent via a Guardian interprocess message), and the

number of messages busy (i.e., the current number of messages that

have been sent via a Guardian interprocess message but have not yet

been replied to) are shown for each Consumer. Messages are queued

to be sent to the Consumer while they are filling with events, or when

there are MAXCONSWRITES messages already outstanding to the

Consumer.

Aside from audit trail timestamp inaccuracies, there are several reasons

COLL STATS may incorrectly report latency when the Collector is

actually caught up they are listed below:

– The ADTSAMPLEDELAY parameter controls how long the

Collector waits after hitting EOF on the audit trail before it goes

and checks for additional audit data. If the STATS command is

issued at this time, the lag (latency) will be increasing however

the Collector is not “behind” – it is waiting for the

ADTSAMPLEDELAY interval to expire before it goes and

checks for more data.

– The audit trail disk processes only flush the audit data to disk

periodically, and the HP ARLIB audit reading routines only

return audit data after it has been flushed to disk. This means

that the disk processes and TMF subsystem may be preventing

HP Shadowbase from reading the audit events, thus

contributing to lag (latency).

– The Collector’s TURBOWAITTIME parameter controls how long

the Collector waits for a buffer to fill before sending it to the

Consumer. In essence, this parameter “holds up” the audit

events from being sent while the buffer fills. This time can add

to the lag times.

Hence, lag values within the ADTSAMPLEDELAY, TURBOWAITTIME,

or TMF flush intervals are not indicative of HP Shadowbase being

“behind”, rather they show the direct effect these parameters have on

overall replication latency.

1 2 ... 97 98 99 100 101 102 103 104 105 106 107 ... 345 346

Comentários a estes Manuais

Sem comentários

HP Integrity NonStop H-Series Guia do Utilizador Página 102

Comentários a estes Manuais

Manuais e produtos relacionados com Software HP Integrity NonStop H-Series