You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@cassandra.apache.org by Richard Hesse <rh...@gmail.com> on 2022/08/01 15:31:59 UTC

Storage load reporting seems off in v4?

After upgrading to version 4.0.3 from 3.11.9 (and running upgradesstables),
we've noticed that the storage load in one of our clusters seems off. That
is, it looks like the counters have wrapped or something like that. Example:

--  Address      Load                  Tokens  Owns (effective)
UN  X.X.X.14   -591611177152 bytes   256     25.6%
UN  X.X.X.24   -487074418074 bytes   256     25.4%
UN  X.X.X.50   121.37 GiB            256     22.3%
UN  X.X.X.25   -839201066679 bytes   256     26.4%
UN  X.X.X.13   -713656184129 bytes   256     25.4%
UN  X.X.X.49   -605482194979 bytes   256     23.5%
UN  X.X.X.15   -736634103675 bytes   256     25.3%
UN  X.X.X.23   -695534031952 bytes   256     25.5%
UN  X.X.X.12   -634839791347 bytes   256     24.9%
UN  X.X.X.16   -780885212531 bytes   256     25.4%
UN  X.X.X.11   -295816469030 bytes   256     26.1%
UN  X.X.X.48   -299002872135 bytes   256     24.2%

Our other clusters went through the same procedure and aren't showing this
behavior. As far as we can tell, things are operating normally outside of
this one reporting issue. Any idea what might be happening here? TIA!

-richard