Red Branch | For Future Reference …
In the update from iDRAC 8 to iDRAC 9 it has been infuriating searching the ‘Storage’ menu only to discover that the storage is managed under ‘Configuration > Storage Configuration’ I’m convinced the user experience could be better! At the very least a link through for configuration.
Journald logs can be cleared by timescale or size as follows. Keep only 5 days worth of logs journalctl –vacuum-time=5d Retain only the past 500 MB: journalctl –vacuum-size=500M
The puppet agent can be disabled from running with:
|
1 |
puppet disable "Some message here" |
And it can be re-enabled with:
|
1 |
puppet enable |
But to check if puppet is currently disabled there is no check command. Instead look for the existence of the agent_disabled.lock file under:
|
1 |
$vardir/state/agent_disabled.lock |
For the systems I was checking this was at:
|
1 |
/opt/puppetlabs/puppet/cache/state/agent_disabled.lock |
I sometimes want to omit directories from my disk usage (du -sh *) checks. This includes virtual files systems such as proc / dev and also mounts to other filesystems (data in the example below). This is what works for me:
|
1 |
du --exclude={proc,sys,dev,data} -sh * |
The equals sign appears to be optional and worked both ways for me…. Read more »
I was checking the memory on approximately 140 servers using ansible:
|
1 |
ansible --user root -k -i servers.list -m shell -a "free -h; mysql" test |
but the output appears over several lines e.g.
|
1 2 3 4 5 |
server103.domain.dev | SUCCESS | rc=0 >> total used free shared buffers cached Mem: 94G 44G 49G 59M 203M 4.6G -/+ buffers/cache: 39G 54G Swap: 0B 0B 0B |
but what I wanted was the server name and the total memory available. So I first grepped the output to remove the lines I didn’t want e.g.
|
1 |
grep -A3 "domain.dev" | egrep -v "^--|Swap|total|buffer" |
leaving me with two lines of… Read more »
In a Cloudera cluster we had a few nodes reporting the following error message in the logs under hadoop-cmf-hdfs-DATANODE-whatever.com.log.out The error included: Datanode denied communication with namenode because the host is not in the include-list: : DatanodeRegistration One solution could be to ‘Refresh Node List’ as shown in the screenshot below. Find it by clicking… Read more »
On the Windows Domain Controller I checked the current NTP settings with:
|
1 |
w32tm /query /configuration |
I tried a few alternative commands and got errors such as “the rpc server is unavailable”. The command that worked was:
|
1 |
w32tm /config /manualpeerlist:time.windows.com /syncfromflags:manual /reliable:yes /update |
After issuing that command I could see the ntp server listed when re-running the command:
|
1 |
w32tm /query /configuration |
As shown below.
I wanted to add some checks to a Windows server and the methods seem to have changed a little over the years. At the moment I have amended the nsclient.ini file on the Windows box with:
|
1 2 |
CheckDisk=1 CheckSystem=1 |
Then from the Nagios box I can check CPU with a threshold as follows:
|
1 |
/usr/lib64/nagios/plugins/check_nrpe -H winserver001.domain.tld --command check_cpu -a "warn=load > 96" "crit=load > 99" "time=5m" |
We can also… Read more »
Using the Hadoop command we can send fairly regular filesystem commands to navigate HDFS. On Cloudera I am sending commands to port 8020 of the active name node. If you don’t have the hadoop command available on your system you need to download hadoop, extract it and add the bin directory to your path. For… Read more »
So working on a system that had incorrect partitioning configured. In /etc/fstab the partitions were configured with:
|
1 2 |
PARTLABEL=data1 /data1 ext4 defaults 0 0 PARTLABEL=data2 /data2 ext4 defaults 0 0 |
but the partitions did not have PARTLABELS set. The LABEL and PARTLABEL can be viewed with:
|
1 |
lsblk -o name,label,partlabel,mountpoint,size,uuid |
To set the label property use:
|
1 |
e2label /dev/sdb1 mylabel |
To set the partlabel on a GPT partition first install gdisk:
|
1 |
yum install gdisk |
Then… Read more »
