HP P2000 G3 monitoring (Zabbix 2.0)

[Emir Imamagic]

Hello,

I attached script and Zabbix template for transfer rates and IOPS for HP P2000 G3 storage system. I used LLD so this solution can be used only in Zabbix 2.0.
For earlier versions it is possible to use statistic gathering & reporting part, but one needs to manually create templates.

Script (hp-msa.pl) uses XML API and it is based on the example provided in "HP P2000 G3 MSA System CLI Reference Guide". Same script is used for LLD and reporting values as "Zabbix trappers". Execute hp-msa.pl to see usage. Unfortunately script does not have decent error/timeout handling, it is in my TODO list.

Template uses LLD to discover all:

• controllers
• virtual disks
• volumes

Discovery item (HP Entities Discovery) uses {HOST.CONN} to connect to the storage system. Statistics gathering item (HP P2000 Sensor) uses {HOST.DNS1} for connecting and value reporting. Make sure that storage host has these two properly configured.

Script execution time is ~10s so one needs to increase Zabbix server config file, e.g.:
Timeout=30

Big thanks to Zabbix team, LLD makes template generation soo much easier.

Best regards,
Emir Imamagic

[elvar]

Wow, your timing is impeccable! I just installed my first P2000 G3 and I came here looking for a template and found this.


Thanks!

[thesubmitter]

Would this work for a g2 msa2312i?

[i.kulagin]

Emir Imamagic
Thanks for template and the script, but can you please provide some kind of guide how to use them? I need to monitor the P2000 G3 FC, but I can't understand how can I do it, using this script, and which services should I configure inside of the P2000 equipment to make monitoring possible.
Thanks for your help.

[scr512]

So i've got most of this working my end but I'm guessing that something isn't quite right.

If I execute the script interactively using the stats option I get a return of 0. I suspect that maybe the P2000s I am running are a version or two behind what you used in testing?

# show inquiry
Inquiry Controller A
--------------------
Management Controller Code Version: L230S025
Management Controller Loader Code Version: 2.5
Storage Controller Code Version: T230P03

[Emir Imamagic]

Hello,

I have exactly the same version so it should work fine.

Is LLD working? Did you get all the controllers, vdisks and volumes configured in Zabbix? Script will report 0 if sending of at least a single value fails.

Unfortunately I didn't have time to do decent debugging so I suggest commenting line:

die "Can not remove file $sendFile!" unless(unlink ($sendFile));

After executing

hp-msa.pl FQDN stats

check content of file "/var/tmp/zabbixSenderHPP2000".

Also check Zabbix server logs for errors.

So i've got most of this working my end but I'm guessing that something isn't quite right.

If I execute the script interactively using the stats option I get a return of 0. I suspect that maybe the P2000s I am running are a version or two behind what you used in testing?

# show inquiry
Inquiry Controller A
--------------------
Management Controller Code Version: L230S025
Management Controller Loader Code Version: 2.5
Storage Controller Code Version: T230P03

[Emir Imamagic]

Before adding anything to Zabbix make sure that the controller has DNS and IP address and that it is accessible from the Zabbix server.

Script uses default username and password (monitor/monitor). If your storage admin changed these parameters, you have to modify them in the hp-msa.pl script, lines:

my $USERNAME = "monitor";
my $PASSWORD = "monitor";

Script uses https, but if that is causing a problem you can switch it to http. Unfortunately I didn't make a parameter for that so you can use sed:

sed -i s/https/http/ hp-msa.pl

Adding script/template to Zabbix:
1. Create host with your controller data (you must IP and DNS address)
2. Import template from the ZIP file
3. On the Zabbix server:
3.1. check the value of ExternalScripts in Zabbix server config file
3.2. put hp-msa.pl in that directory
3.2. make sure that user zabbix has execution permission on hp-msa.pl
4. Link host with the template Template_HP_P2000_G3

Discovery rule has a long delay (3600s) so in order to get LLD items defined you might want to temporarily change it to something like 60.

Wait until items (controllers, vdisks, volumes) appear. Afterwards external item "HP P2000 Sensor" should start reporting values and eventually switch from 0 to 1.

Emir Imamagic
Thanks for template and the script, but can you please provide some kind of guide how to use them? I need to monitor the P2000 G3 FC, but I can't understand how can I do it, using this script, and which services should I configure inside of the P2000 equipment to make monitoring possible.
Thanks for your help.

[Emir Imamagic]

Unfortunately I don't have such storage here. Best way to find out is to give it a try

Would this work for a g2 msa2312i?

[scr512]

Hello,

I have exactly the same version so it should work fine.

Is LLD working? Did you get all the controllers, vdisks and volumes configured in Zabbix? Script will report 0 if sending of at least a single value fails.

Unfortunately I didn't have time to do decent debugging so I suggest commenting line:

die "Can not remove file $sendFile!" unless(unlink ($sendFile));

After executing

hp-msa.pl FQDN stats

check content of file "/var/tmp/zabbixSenderHPP2000".

Also check Zabbix server logs for errors.

So actually the script looks to be working just fine, at least when I run it interactively from CLI I can see a bunch of useful data I would love to ingest.

So, looking at the zabbix_server.log it is obvious that I am doing something wrong in my host configuration.

File does not exist: 500 Can't connect to *UNKNOWN*:443 (Bad hostname '*UNKNOWN*')
 at /usr/local/share/zabbix/externalscripts/hp-msa.pl line 120

I'll poke around and see what I am doing. I will note that when I configured the hosts, I used an SNMP interface, not an agent interface. I do have an IP and DNS name set on the SNMP interface.

[Emir Imamagic]

So actually the script looks to be working just fine, at least when I run it interactively from CLI I can see a bunch of useful data I would love to ingest.

So, looking at the zabbix_server.log it is obvious that I am doing something wrong in my host configuration.

File does not exist: 500 Can't connect to *UNKNOWN*:443 (Bad hostname '*UNKNOWN*')
 at /usr/local/share/zabbix/externalscripts/hp-msa.pl line 120

I'll poke around and see what I am doing. I will note that when I configured the hosts, I used an SNMP interface, not an agent interface. I do have an IP and DNS name set on the SNMP interface.

The tricky part is that I used DNS name for stats. You have to make sure that:
1. DNS attribute is properly defined in "Agent interfaces"
2. DNS can be used to access storage controller and
3. DNS is equivalent to the "Host name" attribute of the host.

[scr512]

The tricky part is that I used DNS name for stats. You have to make sure that:
1. DNS attribute is properly defined in "Agent interfaces"
2. DNS can be used to access storage controller and
3. DNS is equivalent to the "Host name" attribute of the host.

Hmm I seem to be going crazy

I can definitely connect to the storage management interface from my Zabbix box using it's DNS name (SSH)

Yet I am still getting le crazy log messages complaining about "*UNKNOWN*". I'll do some more Googling.


So, does this look sane to you?

[Emir Imamagic]

Hmm I seem to be going crazy

I can definitely connect to the storage management interface from my Zabbix box using it's DNS name (SSH)

Yet I am still getting le crazy log messages complaining about "*UNKNOWN*". I'll do some more Googling.


So, does this look sane to you?

You need to configure add at least one Agent interface and put IP and DNS there. Afterwards you can remove this SNMP interface.

[scr512]

You need to configure add at least one Agent interface and put IP and DNS there. Afterwards you can remove this SNMP interface.

Ok, I have that now setup. I've also configured my zabbix_agentd.conf to have a Timeout=30.

cat /etc/zabbix/zabbix_agentd.conf | grep Timeout
### Option: Timeout
#	Spend no more than Timeout seconds on processing
Timeout=30

So, I seem to be getting a little further along. I am now running into this

1328:20120718:164423.314 item [jckp2k-1a.rac.acl:hp-msa.pl["{HOST.DNS1}","stats"]] became not supported: timeout while executing a shell script
1328:20120718:164443.325 item [jckp2k-1a.rac.acl:hp-msa.pl["{HOST.CONN}"]] became not supported: timeout while executing a shell script
1328:20120718:164528.345 item [jckp2k-1b.rac.acl:hp-msa.pl["{HOST.DNS1}","stats"]] became not supported: timeout while executing a shell script
1326:20120718:164553.362 item [jckp2k-1b.rac.acl:hp-msa.pl["{HOST.CONN}"]] became not supported: timeout while executing a shell script

[Emir Imamagic]

Ok, I have that now setup. I've also configured my zabbix_agentd.conf to have a Timeout=30.

You need to put this variable to Zabbix server config.

Cheers

[scr512]

You need to put this variable to Zabbix server config.

Cheers

Yes, this is set in server config as well... no joy.