Adding a new disk to store logs

[WillemDH]

Hello,

I'm in the process of adding a new disk to my NLS nodes to get some extra storage and wanted to be sure I'm following 'best' practices.
So I added a new 250 GB eager zeroed disk on a separate SCSI controller in VMware.
hen I did

Code: Select all

fdisk -l
fidks /dev/sdb
n
p
1
<enter>
<enter>
w
mkfs.ext4 /dev/sdb1
mkdir /mnt/data
mount /dev/sdb1 /mnt/data

It seems like the disk is correctly mounted. Is my above procedure for adding a new virtual disk plus minus how you guys would add a new disk to a NLS node? I made it ext4 as the sda1 is also ext4. I guess I should add it to /etc/fstab so it stays mounted afetr a reboot?

Then I would just need to follow http://assets.nagios.com/downloads/nagi ... erver.pdf and add /mnt/data to DATA_DIR?

Thanks for confirming or 'suggesting' a better way.

Willem

[jdalrymple]

Process looks sane to me. 2 things I'd point out:

1) Growing the ext4 later down the road is more convoluted if you don't lay it on an LV. It's possible, and even easy, but does require downtime. Since your disk is virtual anyway you'll never need to extend to a new physical disk (unless you run out of room in the datastore and can't relocate that disk), so realistically the difference is mostly aesthetic.

2) Also probably unimportant just because NLS nodes are so easily recycled - but I personally don't know the mechanics of using multiple paths in ElasticSearch. I can't outright think of a situation where it could be an issue, except potentially relocating the data down the road. Like I said though - NLS nodes are easily recyclable so if `mv` is a construct that doesn't work with that conifguration you can just create a new node, join it, then unjoin the old.

Otherwise, it looks like you're on the right track to me.

[WillemDH]

This can be closed. Disk added and extended successfully.