lusty scarab
Hi all, We've recently installed our SNMP monitoring tool onto our NetApp C250 AF Cluster and I've noticed a couple of alerts for the root aggregate being low on disk space (screenshot attached).
Looking deeper into it, there is 159.89GB on the aggr0 root and it has 7.75GB free. Is this normal to use this much space? Would it be something which should be expanded?
Thanks in advance 🙂
granite mulch
Normally aggr0 will not run out of space, because you cannot put any data on it, because it is only for system data... But you can look into a few ways to increase its size... One way would be to add disks to it... My guess is that you have at least 8 disks in your C250, and you might have empty spare partitions that you can add to the aggregate... run this "storage aggregate show-spare-disks" to see if you have any spare partitions... But remember that you cannot remove the partition again... Another way would be to modify the "vol0" so run "set adv" and then "volume show vol0" if you see a lot of empty space here, you can resize "volume resize" or maybe modify it to be thin-provisioned..
cinder sequoia
Yes, it's totally normal for root-aggregates to be 95% "full" since the root-volume inside is thick-privisioned. Please exclude the root-aggrs from monitoring. There is really no use to monitor them.
You could monitor the root-volume size if you want but in general they shouldn't get full.
It could happen if you manually start packet traces for troubleshooting or if core dumps are being created, you could then manually delete them: https://docs.netapp.com/us-en/ontap/system-admin/manage-node-root-volumes-aggregates-task.html#free-up-space-on-a-nodes-root-volume
You can check yourself, while the aggr is 95% used, the vol is only 26% used (in this example):
clusterA::*> aggr show -has-mroot true -fields size,availsize,used,percent-used
aggregate availsize percent-used size usedsize
--------- --------- ------------ ------- --------
n1_root 7.66GB 95% 159.9GB 152.2GB
n2_root 7.66GB 95% 159.9GB 152.2GB
2 entries were displayed.
clusterA::*> vol show -vserver n1,n2 -fields size,available,used,percent-used
vserver volume size available used percent-used
------- ------ ------- --------- ------- ------------
n1 vol0 151.3GB 105.2GB 38.48GB 26%
n2 vol0 151.3GB 105.8GB 37.93GB 26%
2 entries were displayed.
Thanks both, just wanted to ensure this wasn't something which we should be too concurned about as our consultant mentioned it 🙂
lunar swallow
Have you deleted previous ontap images from your upgrades. Could be another reason I think.
hollow hare
yeah, check for older ONTAP images, maybe forgotten packet traces, stuff like that. Also old log files (visible in the SPI) can often go in the hundreds of megabytes or even gigabytes
cinder sequoia
Have you deleted previous ontap images from your upgrades. Could be another reas...
Again: 95% used for the root-AGGRS is totally fine. 95% used for the root-VOLS could lead to problems. But none of the screenshots show the root-vol used-size so we don't know if they are almost full or not.
willow crest
Setting the snap reserve to 0 on the root aggr might help to get it under the alert threshold
granite mulch
Setting the snap reserve to 0 on the root aggr might help to get it under the al...
I am pretty sure they are 0% by default now... unless you are mirroring your aggregates like on MCC... 🙂
We have systems where we have just resized the vol0 volumes a bit smaller... And we have systems that has been expanded over time with new disks, so you typically end up with some spare root-partitions as you expand your aggregates (the disks are automatically partitioned if the raid group you are adding then to, have partitioned disks)... so rather than having a lot of small root-partitions we choose to add them to the aggr0, which might also help just a little bit on performance for the processes using aggr0...