Celerra/Clariion monitoring CPU,Memory,CIFS&NFS activity

[brund]

This is a topic which explains how to monitor Celerra:

On the Control Station

1) Modify the file /etc/init.d/snmd (see model in gzip file )
2) Modify the file /etc/ snmp/snmpd.conf (see model in gzip file )
3) Reload the daemon snmpd (/etc/init.d/snmpd restart)
4) Launch the script Celerra_SNMP_report.ksh ( script in gzip file )

On Cacti

5) import in cacti the template cacti_host_template_celerra.xml
6) copy the file emc_celerra_fs.xml in the directory of cacti /resource/snmp_queries
7 ) in cacti create device with Ip address of CS ( host Template "celerra" )

Show graphs

[chaachee]

Brund;

What version of Celerra code are you using? I've got 5.6 and can't get your Celerra_SNMP_report.ksh script to run:

/outils/system/snmp/CELERRA_SNMP_FS_CONFIG.host_storage02' for reading (No such file or directory)

I did a search of the device and found no files with this name. Any idea why this would be?

[tomer]

Brund;

What version of Celerra code are you using? I've got 5.6 and can't get your Celerra_SNMP_report.ksh script to run:

/outils/system/snmp/CELERRA_SNMP_FS_CONFIG.host_storage02' for reading (No such file or directory)

I did a search of the device and found no files with this name. Any idea why this would be?

there is more tha one error in theese files. i'malso trying to get them running. here's what i found till now

In file Celerra_SNMP_report.ksh

Replace Line 84 (HOSTNAME=`hostname`) with
HOSTNAME="hostname"

Replace Line 105 (INFO=`/nas/bin/server_df server_2 /$FSNAME | tail -1`) with
INFO=`/nas/bin/server_df server_2 $FSNAME | tail -1`

Now the script is running. Have to see if i manage to get the rest working.

[skywalka]

Hey tomer

Did you manage to get it working? If so, what else needed changing?

L.

[skywalka]

I got it all working with Celerra code 5.6 Big thanks to brund & tomer!!!

I didn't need to change line 84 but I did change line 105 as recommended.

[tomer]

hi all

my installation isn't running like i expected. therefore i have some questions and i hope that someone can answer them.

1. do i have to run the Celera_SNMP_report.ksh permanently (lets say with a cronjob every minute) on the celerra control station ?

2. i can see all needed graphs in cacti but none of them has some values in it (so the graphs are empty). issuing an snmpwalk from the cacti host is running. Also the Script Celera_SNMP_report.ksh is running fine. is there something more i can check ?

3. All my scripts running on the Control Station are running in /home/nasadmin/scripts so i would like to change the path /outils/system/snmp to /home/nasadmin/scripts/snmp but i can't figure out where i could set this value.

+ Executing SNMP walk for list of indexes @ '.1.3.6.1.4.1.2021.3400.1.101'
+ Index found at OID: '1.3.6.1.4.1.2021.3400.1.101.1' value: '/bin/cat: /outils/system/snmp/FSIDX/1: No such file or directory'

4. i'm not sure if i understood completly how the moitoring is working. does the Celera_SNMP_report.ksh collect some values and write them into the files. then cacti will connect to the control station and get those values from the files ?

I also looked a little bit further to solve my problem. i remarked that the Poller Cache in Cacti does soemthing like:

ehns700c - CPU Usage - Nice SNMP Version: 2, Community: pf_nas, OID: 1.3.6.1.4.1.2021.3100.1.101.2
RRD: /usr/share/cacti/rra/ehns700c_memfree_4796.rrd

When i try to get the values from the Cacti systems itself:

[root@zfns700c nasadmin]# snmpwalk -v 1 localhost -c public 1.3.6.1.4.1.2021.3100.1.101.2
[root@zfns700c nasadmin]#

so it seems that there is no feedback, like the MIB variable doesn't exist. I guess that this is also the reason why i'm not able to see any graphs, right ? Do you have a hint for me what i will have to change to get some graphs ?


thanks for your feedback

[trevorj]

Does anyone have any new information on these templates?

I recently started working on a Celerra NX4 and I cannot get the .ksh file to run.

[root@xxxx utils]# ./Celerra_SNMP_report.ksh
awk: cmd. line:2: fatal: cannot open file `/outils/system/snmp/CELERRA_SNMP_FS_CONFIG.xxxx' for reading (No such file or directory)
/bin/rm: cannot remove `/outils/system/snmp/FS/NAME': No such file or directory
/bin/rm: cannot remove `/outils/system/snmp/FS/SIZE': No such file or directory
/bin/rm: cannot remove `/outils/system/snmp/FS/CAPA': No such file or directory
cat: /outils/system/snmp/CELERRA_SNMP_FS_CONFIG.xxxx: No such file or directory


[root@xxxx utils]# snmpwalk -v 1 localhost -c public 1.3.6.1.4.1.2021.3100.1.101.2
Timeout: No Response from localhost



Any help would be greatly appreciated.

[skywalka]

you have to create the directories before you can get started:
mkdir -p /outils/system/snmp

then try running the script again, good luck

[trevorj]

Thanks for the response!

I have the script running now, and I can snmpwalk to get correct values. However, the query does not seem to grab all the info necessary to populate the graphs. Did anyone else have this problem?

[rcaston]

While this sounds odd, the more complex filesystem capacity/usage graphs are working fine for me.

But the basic cpu,mem,and cifs/nfs graphs are not, those 3 all just show blank graphs with "-1.#J" for all the values. I checked the snmp polling cache for them and the oids they show are correct, and when queried return proper data.

Curious...

[montjoy]

It looks like you need to create the "$SNMP_HOME/CELERRA_SNMP_FS_CONFIG.hostname" config file. I see there is an example in the .ksh script, but I don't really get how the examples correlate to real Celerra filesystems. Can you give us any more clues? I'm running 5.6.48-7.

[montjoy]

figured it out. the $SNMP_HOME/CELERRA_SNMP_FS_CONFIG.hostname file should have entries like the following ONLY, no comments or other stuff.

1#u01
2#user_home
3#next_filesystem_name


Also, I had to change the "1" to "INFO" for the following in /etc/snmp/snmpd.conf to get the CPU/mem/NFS/CIFS info

exec .1.3.6.1.4.1.2021.3100.1 SYSTEM /bin/cat /outils/system/snmp/SYS/INFO
exec .1.3.6.1.4.1.2021.3200.1 CIFS /bin/cat /outils/system/snmp/CIFS/INFO
exec .1.3.6.1.4.1.2021.3300.1 NFS /bin/cat /outils/system/snmp/NFS/INFO

[beermann]

Hi, I am trying this on my celerra ver. 6.0.36-4 but the command server_nfsstat has been deprecated and it tells me that i should use server_stats instead, but how do i get the same stats with this command? and also i do not have any memory graph in cacti althoug the system is providing the data.

[xolo]

useful for me,thanks