EMC monitoring using check_emc_clariion.pl

[Box293]

Can you let please execute the following command and let us know the results. This is what the plugin is executing to generate it's output, so I'll be able to compare them to see what is happening.

Code: Select all

/opt/Navisphere/bin/naviseccli -User nagios -Password password -Scope 0 -h emc_host getlun 407 -name -type -state -usage -capacity -drivetype -prb -bind

Make sure you replace the emc_address and password with the correct values.

[slansing]

Can you show us the actual amount of used space from that system? Without using a plugin? Looks like it is the perfdata that is showing up as 0% free, and that is likely also what the plugin is seeing when it tells you that it is lower than the threshold you specified.

[lannister]

Box293-:
I did execute the command -: /opt/Navisphere/bin/naviseccli -User nagios -Password password -Scope 0 -h emc_host getlun 407 -name -type -state -usage -capacity -drivetype -prb -bind
Name DEMO_TEST
RAID Type: RAID5
State: Bound
LUN Capacity(Megabytes): 10240
LUN Capacity(Blocks): 20971520
Drive Type: SAS
Prct Rebuilt: 100
Prct Bound: 100

Its the similar output it gives for all the other luns.

slansing -:

The actual amount of space is its almost 90 % free aprt from that i have also tried on other luns and raid groups , it gives the same output.

Is it becoz the lun allocated from emc to the host will always show full as in suppose we assign 100 Gb lun so it will show 100 Gb full as the lun is being assigned to host and its not able to pull how much its been utilized by the host.

- Lan

[sreinhardt]

Considering we do not know how your particular device is configured, and do not have one in-house to test against, it's very hard to say how this plugin should be responding. I can say that considering the output, this looks very promising that it is in fact reading your device's snmp daemon and reporting back the information. Your logic for why the luns are showing as they are, does seem mostly sane, although I would have expected the device to know what space is actually available and free.

[Box293]

From looking at your output, it would appear as though the plugin is correctly reporting the amount used.

Prct Bound: 100

This value states the LUN is 100% used.

If your host is an ESX(i) host, then it will report back to the SAN that it has consumed all of the space, even though that is not the case when looking at the LUN via the vSphere client. This is just how it is. I have seen in other SAN's (like EqualLogic) where a LUN/VOLUME can be thin provisioned, meaning that the ESX(i) host correctly reports back to the SAN the actual space that has been used. This all depends on the SAN's capabilities.

I recommend monitoring the LUN via the HOST that is connect to the LUN, this way you will get the utilization information you are after. NOTE: It is still important to monitor the LUN via the EMC SAN as this will report if a disk failed in a raid group etc.

[lannister]

Hi ,

Okay I totally understand the logic explained for the luns,but then for Raid groups it shouldn't show the same output as that of luns.
Even for Raid Groups it shows 100 % full , though it isnt.

Thanks for your time and patience.

- Lan

[slansing]

@lannnister,

I would tend to agree, as Sreinhardt mentioned we have nothing to test this against in house however I believe Box293 has the means to answer your questions accurately, unfortunately he works virtually opposite hours than us.

[Box293]

It's been over a year since I've worked on a EMC SAN however my understanding of a Raid group is as follows:

• A raid group is a collection of physical disks configured with a specific raid type
  For example, 5 x 1TB drives in a RAID 5 is approximately 4TB
  A LUN is created from the raid group and this is what the hosts connect to
  When a LUN is created, the amount of space allocated to the lun is "used" in the raid group
  So if you created a 2TB LUN, then the RAID group would have 2TB used (50%)
  So if you had just 1 x 4TB LUN created then the RAID group would be 100% used
  The term Full or Used probably isn't correct here, it's more of an allocation

If you execute the following command, it will report some stats about the raid group:

Code: Select all

/opt/Navisphere/bin/naviseccli -User nagios -Password password -Scope 0 -h emc_host getrg group_id -type -tcap -ucap -prcntdf -prcntex

Make sure you replace the emc_address, password and group_id with the correct values.

Specifically there should be a line Free Capacity (Blocks,non-contiguous):. If this value is 0, then the raid group is completely allocated.

I'm pretty sure I'm right but it's been a while since I've used an EMC SAN so my memory is a little foggy ... and newer firmwares could have added configurations I'm not aware of.

[lannister]

Yup , absolutely right .

Below is the output -:
RaidGroup ID: 20
RaidGroup Type: rx
Raw Capacity (Blocks): 30744870912
Logical Capacity (Blocks): 23058653184
Free Capacity (Blocks,non-contiguous): 0
Percent defragmented: 100
Percent expanded: N/A

Is there a way to find out for pools and file systems as well using the same plugin ?
I know im asking for a bit too much

Thank you

- Lan

[tmcdonald]

I personally am not too familiar with this plugin, but when Box gets back online he might have more for you.