Page 1 of 1
Snapshot > Repository - Sizing question
Posted: Fri Feb 22, 2019 4:46 pm
by SteveBeauchemin
I am taking over support for the Nagios Log Server we run. Someone retired and I get the support task.
I see that there are no repository configured for snapshots. I have a 4 node cluster and understand the CIFS/NFS issue. All my Nagios users and groups UID and GID are consistent across my systems so I am not worried about that.
What I am concerned with is the sizing. I cannot grok the size requirements from the documentation I have read.
I am getting the Log Server healthy. It has been ignored for a long time but I plan to use it heavily now.
I just need to know clearly how much space to reserve for the common location where the repository data will go.
I have 4 servers with 800GB drive space allocated on each.
So, doing the match, 3.2 TB total.
Not much is used right now. Less than 6GB in use at this moment.
Please advise.
Thanks
Steve B
Re: Snapshot > Repository - Sizing question
Posted: Mon Feb 25, 2019 10:21 am
by cdienger
The repos will hold a complete copy of the indices that are snapshotted and will need to be sized according to how many indices you'll want to keep on the repo and their sizes. You can get an idea of the space required by going to Admin > System > Cluster Status and looking at the primary sizes.
Re: Snapshot > Repository - Sizing question
Posted: Mon Feb 25, 2019 11:12 am
by SteveBeauchemin
Thanks you. That is good information.
Final question on this topic: I have a 4 node NLS cluster, should the snapshots go to a non-NLS system as a best practice?
Thanks.
Steve B
Re: Snapshot > Repository - Sizing question
Posted: Mon Feb 25, 2019 11:15 am
by cdienger
Yes, snapshots should go to another remote server and not a NLS node.
Re: Snapshot > Repository - Sizing question
Posted: Mon Feb 25, 2019 2:39 pm
by SteveBeauchemin
okay, I think I'm good.
Nagios user exists on all 5 systems with the same UID and GID.
On all 5 systems:
Code: Select all
mkdir /usr/local/backups
chown nagios:nagios /usr/local/backups
Then on the non-NLS host I added this line to the /etc/exports file
Code: Select all
/usr/local/backups <IP1>(rw,root_squash,sync) <IP2>(rw,root_squash,sync) <IP3>(rw,root_squash,sync) <IP4>(rw,root_squash,sync)
Where IP1, IP2, etc are the IP of the NLS hosts. That way I can enforce security on the location.
Next I exported the file system.
Then, on all 4 NLS hosts I did this: (maybe overkill, but I don't care)
Make sure NFS is running so I can import a file system
Code: Select all
systemctl enable rpcbind
systemctl enable nfs-server
systemctl enable nfs-lock
systemctl enable nfs-idmap
systemctl start rpcbind
systemctl start nfs-server
systemctl start nfs-lock
systemctl start nfs-idmap
Edit the /etc/fstab file and add one line:
Code: Select all
<Exported-IP-of-disk-location>:/usr/local/backups /usr/local/backups nfs defaults 0 0
Then I mount it.
Code: Select all
mount <Exported-IP-of-disk-location>:/usr/local/backups
And tested, as the nagios user, by adding files from each NLS host. I could see and delete from each system.
From the NLS GUI, I made up a Repository name, and added /usr/local/backups
Looks like it worked. I just wanted to be really clear on how I did this work. Good docs help everyone save time.
Thanks
Steve B
(You can close this)
Re: Snapshot > Repository - Sizing question
Posted: Mon Feb 25, 2019 3:38 pm
by cdienger
Thanks for sharing your steps!