[CASSANDRA-8227] Negative load after truncating tables

• Type: Bug
• Status: Resolved
• Priority: Normal
• Resolution: Duplicate
• Component/s: Legacy/Tools

Environment

2.1.1, EC2MultiRegionSnitch

• Severity: Normal
• Since Version: 2.1.1
• Assignee: Unassigned
• Reporter: Juho Mäkinen
• Created: 31/Oct/14 07:52
• Updated: 16/Apr/19 09:31
• Resolved: 31/Oct/14 07:54

Description

After two TRUNCATE command the nodetool status started reporting negative loads on cluster. This bug is closely related to CASSANDRA-7239 which was fixed to 2.1.1, but it seems that probably some edge case was missed regarding to table truncation.

– Address Load Tokens Owns Host ID Rack
UN 54.88.4.xxx -5079005841 bytes 256 ? 17c00df4-ae94-47cb-ab8a-7e97bbfb1c20 1e
UN 54.85.254.xxx -2215232194 bytes 256 ? 0ad0a2d1-0384-4f20-abbe-7834b24b5446 1b
UN 54.209.26.xxx -6804377578 bytes 256 ? 7da4745e-131d-406f-b446-9483f8aa6307 1d
UN 54.86.79.xxx 201.83 MB 256 ? 22d1d5a2-50aa-45c3-bdf6-d15f943d0303 1e
UN 54.86.216.xxx 35.03 GB 256 ? 394b963b-f993-4cc3-be62-2dfb4a46653b 1d
UN 54.84.212.xxx -2226072171 bytes 256 ? e35f5222-72e6-4306-b0df-e36501461e00 1d
UN 54.84.21.xxx -1380645061 bytes 256 ? 842a60a5-a150-4b93-a72c-61720893c363 1b
UN 54.209.8.xxx -7546675109 bytes 256 ? ff48b58e-32ad-4572-a58c-1aadee96b0d0 1b
UN 54.84.135.xxx 213.96 MB 256 ? b23422a7-e141-48ad-956e-28239e170213 1e

Issue Links

• duplicates: CASSANDRA-8205 ColumnFamilyMetrics#totalDiskSpaceUsed gets wrong value when SSTable is deleted