Backport #15652
closedhammer: check-head-obj-locator fix failing & bucket non-deletable
Description
We've got two buckets where the Hammer fixup check-head-obj-locator claims it needs fixing, but it can't fix it, and bucket rm fails to remove the bucket.
- radosgw-admin bucket check --check-head-obj-locator --bucket="jenkins-jucxk1jxn18013k9jftva-202" --fix
ERROR: fix_head_object_locator() returned ret=-22016-04-28 18:43:18.428011 7f8f011d1780 -1 ERROR: ioctx.operate(oid=default.192572025.16587____) returned ret=-2
{
"bucket": "jenkins-jucxk1jxn18013k9jftva-202",
"check_objects": [
{
"key": {
"type": "head",
"name": "__",
"instance": ""
},
"oid": "default.192572025.16587____",
"locator": "default.192572025.16587___",
"needs_fixing": true,
"status": "needs_fixing"
},
{
"key": {
"type": "tail",
"name": "__",
"instance": ""
},
"needs_fixing": false,
"status": "ok"
}
]
}
- radosgw-admin --bucket jenkins-jucxk1jxn18013k9jftva-202 bucket stats
{
"bucket": "jenkins-jucxk1jxn18013k9jftva-202",
"pool": ".rgw.data.1",
"index_pool": ".rgw.data.1",
"id": "default.192572025.16587",
"marker": "default.192572025.16587",
"owner": "jenkins-congress-main",
"ver": "0#28",
"master_ver": "0#0",
"mtime": "2016-02-14 10:20:46.000000",
"max_marker": "0#",
"usage": {
"rgw.main": {
"size_kb": 0,
"size_kb_actual": 0,
"num_objects": 1
}
},
"bucket_quota": {
"enabled": false,
"max_size_kb": -1,
"max_objects": -1
}
}
- radosgw-admin --bucket jenkins-e6fttd5hpwcaof3izpevp-203 bucket check --check-head-obj-locator --fix
ERROR: fix_head_object_locator() returned ret=-22016-04-28 18:46:36.330858 7fcf699f1780 -1 ERROR: ioctx.operate(oid=default.192572025.7454____) returned ret=-2
{
"bucket": "jenkins-e6fttd5hpwcaof3izpevp-203",
"check_objects": [
{
"key": {
"type": "head",
"name": "__",
"instance": ""
},
"oid": "default.192572025.7454____",
"locator": "default.192572025.7454___",
"needs_fixing": true,
"status": "needs_fixing"
},
{
"key": {
"type": "tail",
"name": "__",
"instance": ""
},
"needs_fixing": false,
"status": "ok"
}
]
}
- radosgw-admin --bucket jenkins-e6fttd5hpwcaof3izpevp-203 bucket stats
{
"bucket": "jenkins-e6fttd5hpwcaof3izpevp-203",
"pool": ".rgw.data.1",
"index_pool": ".rgw.data.1",
"id": "default.192572025.7454",
"marker": "default.192572025.7454",
"owner": "jenkins-congress-main",
"ver": "0#27",
"master_ver": "0#0",
"mtime": "2016-02-11 23:49:54.000000",
"max_marker": "0#",
"usage": {
"rgw.main": {
"size_kb": 0,
"size_kb_actual": 0,
"num_objects": 1
}
},
"bucket_quota": {
"enabled": false,
"max_size_kb": -1,
"max_objects": -1
}
}
Updated by Nathan Cutler almost 8 years ago
- Tracker changed from Bug to Backport
- Release set to hammer
Updated by Nathan Cutler over 7 years ago
- Status changed from New to Need More Info
- Assignee set to Robin Johnson
Robin, is there a master PR we can cherry-pick, or is this a hammer-only bugfix?
Updated by Robin Johnson over 7 years ago
I don't have any buckets immediately that still exhibit this problem in Jewel, sorry :-(.
Updated by Nathan Cutler about 7 years ago
- Status changed from Need More Info to Closed
Closing because hammer backports are focusing on bugs that prevent users from upgrading their hammer clusters to jewel.