bash-3.00# prctl 5517
process: 5517: db2sysc

..

project.max-sem-ids
privileged      128       -   deny                                 -
system          16.8M     max   deny                                 -

..

Changed it to 4000 and it is working fine:

prctl -n project.max-sem-ids -r -v 4000 5517

This is vmware esx3.5. Today I found out the same mistake was made, with vmware snapshot not deleted for a very long time. Tried to delete the snapshot, but it timed out and after rebooting guest, I hit the same error as this one: http://sysadminhell.blogspot.com/2008/03/couple-of-vmware-issues.html. I followed the trick and killed vm, then started it up again. Got a snapshot named “Console Helper-1″. I tried to delete it again, it timed out again. It seems there is no way to see its progressing rate, the only way from vmware is:

watch -d “ls -touh –full-time *.vmdk” to just tell file modification timestamp changes for delta files.

(http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1007566)

But it seems it is in the process of deleting snapshot still.

StoreConfigs not supported without ActiveRecord 2.1 or higher

Turns out it could only work with ActiveRecord major version 2. Need to downgrade activerecord on this Ubuntu 10.04:

gem list:

activerecord (3.0.1)

gem install activerecord –version 2.3.5

gem list:

activerecord (3.0.1, 2.3.5)

# gem uninstall activerecord –version 3.0.1
Successfully uninstalled activerecord-3.0.1

activerecord (2.3.5)

# /etc/init.d/puppetmaster restart
* Restarting puppet configuration management tool master server         [ OK ]

Verify ok:

# racadm -r idrac_ip_address -u root -p calvin getconfig -g cfgUserAdmin -i1
Security Alert: Certificate is invalid – self signed certificate
Continuing execution. Use -S option for racadm to stop execution on certificate-related errors.
# cfgUserAdminIndex=1
cfgUserAdminUserName=
# cfgUserAdminPassword=******** (Write-Only)
cfgUserAdminEnable=0
cfgUserAdminPrivilege=0×00000000
cfgUserAdminIpmiLanPrivilege=4
cfgUserAdminIpmiSerialPrivilege=15
cfgUserAdminSolEnable=0

/dev/sda3 (Dell R910 local SAS drive)

Read performance:

# dd iflag=direct if=/dev/md0 of=/dev/null bs=1M count=10000

# echo “scale=3;640/136″|bc -l

4.705

Write performance:

# dd if=/dev/zero of=/FusionMount/wuhai.txt bs=1M count=1000 conv=fdatasync
1000+0 records in
1000+0 records out
1048576000 bytes (1.0 GB) copied, 3.35998 seconds, 312 MB/s

# dd if=/dev/zero of=/tmp/wuhai.txt bs=1M count=1000 conv=fdatasync
1000+0 records in
1000+0 records out
1048576000 bytes (1.0 GB) copied, 9.21373 seconds, 114 MB/s

$ echo “scale=3;312/114″|bc -l
2.736

FusionIO cards in RAID0:

# dd iflag=direct if=/dev/md0 of=/dev/null bs=1M count=10000

10000+0 records in

10000+0 records out

10485760000 bytes (10 GB) copied, 10.1195 seconds, 1.0 GB/s

/proc/sys/dev/raid/speed_limit_min is 1000, and /proc/sys/dev/raid/speed_limit_max is 200000.

After changing /proc/sys/dev/raid/speed_limit_max to 800000, noticed RAID1 mirror resync speed goes up to 516M/sec, much faster than the default one, which is around 200M/sec.

# cat /etc/profile.d/facter.sh
export FACTERLIB=/var/lib/puppet/lib

2. on puppet master server:

$ cat /etc/puppet/modules/gateway/lib/facter/gateway.rb
# gateway.rb

Facter.add(“gateway”) do
setcode do
%x{ip route | grep default | awk ‘{print $3}’}
end
end

3. Ensure puppet client has:

[main]

factpath=$vardir/lib/facter
pluginsync=true

in /etc/puppet/puppet.conf

Then facter | grep gateway would provide gateway IP as a fact.