[prev in list] [next in list] [prev in thread] [next in thread] 

List:       dlm-devel
Subject:    [Dlm-devel] status of 2 node bring up
From:       Alan Jones <alan.jones () sun ! com>
Date:       2001-04-10 1:57:40
[Download RAW message or body]

All,

With the hb patch for giving the kernel a full membership list with
each change, and my solution #2 for running the scheduler code from
clm_master_loop() directly out of each network message receive routine:

- joining node send CLMR_NEW to the already running node

- the already running node fail the test at the bottom of
  rl_enum_counted()

The numbers are far appart:

clm_rldb.c rl_enum_counted 925 ctd_slot 0 count 32 rldbtabsize 100000
clm_rldb.c rl_enum_counted 966 last 32 rldbtabsize 100000

It appears to expect many such messages to complete a 100K item table.
But only one message is generated.

Any suggestions?

Alan

[prev in list] [next in list] [prev in thread] [next in thread]