Bug 1243537 - [rhsc] trends tab of particular host shows two graphs with the same label "Mount Path: /"
Summary: [rhsc] trends tab of particular host shows two graphs with the same label "Mo...
Keywords:
Status: CLOSED DUPLICATE of bug 1167425
Alias: None
Product: Red Hat Gluster Storage
Classification: Red Hat Storage
Component: rhsc
Version: rhgs-3.1
Hardware: Unspecified
OS: Unspecified
unspecified
low
Target Milestone: ---
: ---
Assignee: Ramesh N
QA Contact: RHS-C QE
URL:
Whiteboard:
Depends On:
Blocks: 1216951
TreeView+ depends on / blocked
 
Reported: 2015-07-15 17:15 UTC by Martin Bukatovic
Modified: 2016-04-06 06:44 UTC (History)
9 users (show)

Fixed In Version:
Doc Type: Known Issue
Doc Text:
Labels do not show enough information for the Graphs shown on the Trends tab. When you select a host in the system tree and switch to the Trends tab, you will see two graphs for the mount point '/': one graph for the total space used and another for the inodes used on the disk. Workaround: The graph with y axis legend as %(Total: ** GiB/Tib) is the graph for total space used. The graph with y axis legend as %(Total: <number>) is the graph for inode usage.
Clone Of:
Environment:
Last Closed: 2016-04-06 06:44:07 UTC
Embargoed:


Attachments (Terms of Use)
screenshot of trends tab for selected host (208.64 KB, image/png)
2015-07-15 17:15 UTC, Martin Bukatovic
no flags Details
nagios Disk Utilization page for given host (281.72 KB, image/png)
2015-07-17 07:37 UTC, Martin Bukatovic
no flags Details

Description Martin Bukatovic 2015-07-15 17:15:59 UTC
Created attachment 1052399 [details]
screenshot of trends tab for selected host

Description of problem
======================

When I click on "Trends" tab of particular host from nagios enabled cluster,
I see 2 grapsh with the same label: "Mount Path: /"

Version-Release number of selected component (if applicable)
============================================================

Console server packages:

on RHEL 6
rhsc-3.1.0-0.62.el6.noarch
nagios-3.5.1-6.el6.x86_64

Storage servers packages:

on RHEL 6
gluster-nagios-common-0.2.0-1.el6rhs.noarch
nrpe-2.15-4.1.el6rhs.x86_64

Steps to Reproduce
==================

1. Create cluster using gluster storage console
2. Enable nagios monitoring for this cluster
3. Select on host from the cluster and click on it's "Trends" tab

Actual results
==============

Multiple graphs are presented, but 2 of them have the same label:
"Mount Path: /".

See attached screenshot.

When I look at the output of df on the selected host:

~~~
# df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/vda2              13G  2.0G   11G  17% /
tmpfs                 939M     0  939M   0% /dev/shm
/dev/mapper/vg--brick1-brick1
                       99G   54G   46G  54% /rhgs/brick1
~~~

I see that the 2nd graph from the "Trends" tab is for tmpfs /dev/shm
filesystem.

Expected results
================

Label for disk usage graphs on host should be more descriptive, probably
including the device of the filesystem.

or

Remove tmpfs filesystems from this tab.

Comment 1 RamaKasturi 2015-07-15 18:29:35 UTC
Hi Martin,

  We already have a bug for this and please take a look at the one below:

https://bugzilla.redhat.com/show_bug.cgi?id=1167425#c4

Thanks
kasturi

Comment 2 Ramesh N 2015-07-16 06:56:00 UTC
Can we close this as duplicate of other bug bz#1167425#c4 mentioned in comment 1?

Comment 3 Martin Bukatovic 2015-07-16 08:16:05 UTC
It depends on how you would like to solve the issue:

If you are going to fix the label so that it will contain block device name,
feel free to close this as a duplicate of bz#1167425#c4.

But if you are going to remove graph for /dev/shm (tmpfs) from host's Trends
tab, it would make more sense to let it open.

Comment 4 Ramesh N 2015-07-17 05:18:21 UTC
(In reply to Martin Bukatovic from comment #3)
> It depends on how you would like to solve the issue:
> 
> If you are going to fix the label so that it will contain block device name,
> feel free to close this as a duplicate of bz#1167425#c4.
> 

Sorry I didn't see the /tmpfs part. Actually we are monitoring multiple aspects of the disk like space used, inodes used, thin pool used (if its a thin lv). So adding a proper label for the graph should fix this issue. 

> But if you are going to remove graph for /dev/shm (tmpfs) from host's Trends
> tab, it would make more sense to let it open.

I don't think we are monitoring /dev/shm in nagios. So you won't find any graph for this device.

Comment 5 Martin Bukatovic 2015-07-17 07:37:57 UTC
Created attachment 1052997 [details]
nagios  Disk Utilization page for given host

(In reply to Ramesh N from comment #4)
> I don't think we are monitoring /dev/shm in nagios. So you won't find any
> graph for this device.

It seems that we are. I rechecked nagios web interface and looked at
graphs for on Disk Utilization given host, see the attachment. Monitoring
has been configured using rhsc-setup and configure-gluster-nagios scripts.

But it's not immediately apparent which graph it is as all share the same
label - "Mount Path: /". So it seems that the issue with labels makes this
issue (monitoring /dev/shm) harder to find.

Comment 6 Ramesh N 2015-07-17 08:10:46 UTC
(In reply to Martin Bukatovic from comment #5)
> Created attachment 1052997 [details]
> nagios  Disk Utilization page for given host
> 
> (In reply to Ramesh N from comment #4)
> > I don't think we are monitoring /dev/shm in nagios. So you won't find any
> > graph for this device.
> 
> It seems that we are. I rechecked nagios web interface and looked at
> graphs for on Disk Utilization given host, see the attachment. Monitoring
> has been configured using rhsc-setup and configure-gluster-nagios scripts.
> 

Labels are available in Nagios UI. You can see the text 'Datasource: Disk Utilization(<space/inode>)' in top left corner of all the graphs to see whats the graph is about. From the attachment, I can see graphs for the mount point '/'. So there is no question of '/dev/shm'. 

> But it's not immediately apparent which graph it is as all share the same
> label - "Mount Path: /". So it seems that the issue with labels makes this
> issue (monitoring /dev/shm) harder to find.

Comment 7 Martin Bukatovic 2015-07-17 08:22:10 UTC
(In reply to Ramesh N from comment #6)
> Labels are available in Nagios UI. You can see the text 'Datasource: Disk
> Utilization(<space/inode>)' in top left corner of all the graphs to see
> whats the graph is about. From the attachment, I can see graphs for the
> mount point '/'. So there is no question of '/dev/shm'.

Yes, you are right. The graph is actually for inode utilization. So this can
be closed as duplicate as suggested in comment 2 after all.

Comment 8 Ramesh N 2015-07-24 11:08:23 UTC
doc text looks good. Approved.

Comment 9 Sahina Bose 2016-04-06 06:44:07 UTC

*** This bug has been marked as a duplicate of bug 1167425 ***


Note You need to log in before you can comment on or make changes to this bug.