On 15.03.2023 12:51, Matthias Heiserer wrote:
On 15.03.2023 12:10, Dominik Csapak wrote:
hi not sure we would want to do it this way, since
this will only work for the node where the new pve-cluster
is installed, the other nodes will not be able to update their
local database with the info from
Hi,
> > > +++ b/debian/pve-cluster.postinst
> > > @@ -0,0 +1,6 @@
> > > +#!/bin/bash
> > > +
> > > +RRD_DB="/var/lib/rrdcached/db/pve2-node/`hostname`"
> > > +if [ -f $RRD_DB ] && [ `rrdtool info $RRD_DB | grep -q arcsize`
> > > ]; then
> > > + rrdtool tune $RRD_DB 'DS:arcsize:GAUGE:120:0
Hi,
AFAIR, we have discussed about this last year (for my need to add new
counters for DRS, like pressure io/ram/disk counters),
and we have talked about 1 metric = 1rrd. (but I'm not sure about
read/write performance). No solution was chosen.
Le mercredi 15 mars 2023 à 12:10 +0100, Dominik Cs
On 15.03.2023 12:10, Dominik Csapak wrote:
hi not sure we would want to do it this way, since
this will only work for the node where the new pve-cluster
is installed, the other nodes will not be able to update their
local database with the info from the new nodes
AFAIR, the way we dealt with rrd
hi not sure we would want to do it this way, since
this will only work for the node where the new pve-cluster
is installed, the other nodes will not be able to update their
local database with the info from the new nodes
AFAIR, the way we dealt with rrd updates in the past is
that we introduced a
To modify the existing rrd store, rrdtool is required
Signed-off-by: Matthias Heiserer
---
data/src/status.c | 1 +
debian/control | 1 +
debian/pve-cluster.postinst | 6 ++
3 files changed, 8 insertions(+)
create mode 100644 debian/pve-cluster.postinst
diff --git a/