Re: [Sks-devel] keyserver.linux.it is changing IP

[Jeff Johnson]

On Jan 28, 2011, at 2:41 AM, Marco Nenciarini wrote:
> 
> Do someone see how these difference can lead to a PTree corruption
> during (at least) the pbuild run?
> 

I can't debug from the reported symptoms. I have seen many
PTree failures over the last year setting up multiple
SKS 1.1.1 servers.

My setup is mostly on Mac OS X or CentOS, all of ppc64, ix86 and x86_64
and always using the latest version of Berkeley DB (5.1.19 currently).

The incidence of PTree issues diminished as rapidly as
getting an estimate for the Berkeley DB tunables to
an adequate "working set".

This is the PTree/DB_CONFIG currently in use here (the Locking
tunables are most important):

set_lg_dir              ./log
set_tmp_dir             ./tmp

set_flags               DB_LOG_AUTOREMOVE

# -- thread_count must be >= 8
#set_thread_count        64

# ================ Logging
set_lg_regionmax        1048576
set_lg_max              10485760
set_lg_bsize            2097152

# ================ Memory Pool
##set_cachesize         0 67108864 4
set_cachesize           0 134217728 1
set_mp_mmapsize         268435456

# ================ Locking
##mutex_set_max         8192
set_lk_max_locks        32768
set_lk_max_lockers      32768
set_lk_max_objects      32768
set_lk_detect           DB_LOCK_DEFAULT

# ================ Replication

Hmmm ... the usual rule I use for scaling would
have
        mutex_set_max   327680
i.e. ~10x the size of the other tunables.

hth

73 de Jeff

smime.p7s
Description: S/MIME cryptographic signature