On 22/12/2020 04:52, Mohan Nagandlla wrote:
HI team I am using the Prometheus instance having the more than 1
replica, When replica as 1 there is no wal corruption in data
directory and now for the sake of zero down time updates for instance
I make the replicas count as 2 the instance is up now but the wal
corruptions are happening the logs are below
|level=error ts=2020-12-22T04:36:00.860Z caller=scrape.go:1076
component="scrape manager" scrape_pool=depl/node-exporter/0
target=http://x.x.x.x:9100/metrics msg="Scrape commit failed"
err="write to WAL: log samples: write /prometheus/wal/00000003: stale
NFS file handle" level=error ts=2020-12-22T04:36:00.862Z
caller=scrape.go:1076 component="scrape manager"
scrape_pool=depl/prometheus-kubelet/0
target=https://x.x.x.x:10250/metrics msg="Scrape commit failed"
err="write to WAL: log samples: write /prometheus/wal/00000003: stale
NFS file handle" level=error ts=2020-12-22T04:36:00.881Z
caller=scrape.go:1076 component="scrape manager"
scrape_pool=depl/prometheus-kubelet/0
target=https://x.x.x.x:10250/metrics msg="Scrape commit failed"
err="write to WAL: log samples: write /prometheus/wal/00000003: stale
NFS file handle" level=error ts=2020-12-22T04:36:00.898Z
caller=scrape.go:1076 component="scrape manager"
scrape_pool=depl/prometheus-kubelet/0
target=https://x.x.x.x:10250/metrics msg="Scrape commit failed"
err="write to WAL: log samples: write /prometheus/wal/00000003: stale
NFS file handle" level=error ts=2020-12-22T04:36:00.970Z
caller=scrape.go:1076 component="scrape manager"
scrape_pool=depl/node-exporter/0 target=http://x.x.x.x:9100/metrics
msg="Scrape commit failed" err="write to WAL: log samples: write
/prometheus/wal/00000003: stale NFS file handle" |
Getting more logs like this at one replica no errors but when i am
using the more than 1 replica getting above errors.
Or is there any other way for prometheus zero down time and why does i
am getting this errors but if i used the replicas as 1 there is no
errors in data directory this is happening more than 1 replica
Prometheus must not share a data directory with another running
instance, as you will see data corruption. Each Prometheus instance must
have a unique data directory. Additionally NFS isn't supported, so you
should use a local hard drive/EBS volume.
--
You received this message because you are subscribed to the Google Groups
"Prometheus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to prometheus-users+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/prometheus-users/fb9bdbaa-dbd3-cf5a-3af8-1205b34dc05e%40Jahingo.com.