Most of what you describe sounds pretty DSE specific so I suspect your best
source for answers will be datastax
There were bugs in some versions of Cassandra that caused corruption during a
few milliseconds race of ALTER TABLE, and if you’re not using compression with
CRC checking enabled, it’s possible a bad disk or bit flip could have corrupted
some of your data, but hard to say much beyond that.
--
Jeff Jirsa
> On May 30, 2018, at 3:42 PM, Charulata Sharma (charshar) <chars...@cisco.com>
> wrote:
>
> Hi,
>
> I am observing a very strange behavior in our cluster. Metadata is being
> prefixed in some rows.
> This metadata cannot be sent by application primarily because application
> writing to C* will not have this data,
> and also applications use custom Java objects and this metadata doesn’t fall
> in that category.
>
> I am suspecting that this is being added during multi data center replication
> between analytics and transactions cluster.
> Does anyone have any idea on this. I found no clue online also.
>
> This is the metadata: The nodes mentioned here are the analytics node and the
> data directory is from DSEFS. We are experiencing
> Some corruption in dsefs and we understood that this is because of the
> version we are on (DSE 5.1.5), so I suspect this could be related.
>
> Thanks,
> Charu
>
>
> {
> "privateAddress": "333.33.333.333",
> "lastUpdate": "2018-04-25 05:05:40.665+0000",
> "readOnly": false,
> "up": true,
> "host": "cssdb-prd-07",
> "publicAddress": "333.33.333.333",
> "privatePort": 5599,
> "storageWeight": 1.0,
> "minFreeSpace": 5368709120,
> "publicPort": 5598,
> "dataCenter": "DC1-RPTG",
> "version": 2,
> "locationId": "498ff9f4-a989-461a-8ce7-17c418006666",
> "rack": "RACK1",
> "estUsedSpace": 29771,
> "nodeId": "97494bd1-e783-4a8b-b180-a1946defc7cc",
> "estFreeSpace": 726682292224,
> "directory": "/cassandra/data/ccrcprd-cluster/data5/dsefs/data"
> }
