Bug #55778
closedclient: choose auth MDS for getxattr with the Xs caps
0%
Description
None of the PRs in the branch seem to be causing this and its not easily reproducible.
pjd fails in:
2022-04-13T04:59:41.926 INFO:tasks.workunit.client.0.smithi167.stderr:+ ntest=80 2022-04-13T04:59:41.926 INFO:tasks.workunit.client.0.smithi167.stderr:+ expect u::r--,u:root:r-x,u:65531:rwx,g::r--,g:root:rw-,g:65532:rwx,m::rwx,o::r-- getfacl fstest_73fed23d6d34bccdc1d73b05e3d22103/fstest_da3b235a078085ef61f2cae3419311a2 access 2022-04-13T04:59:41.926 INFO:tasks.workunit.client.0.smithi167.stderr:+ e=u::r--,u:root:r-x,u:65531:rwx,g::r--,g:root:rw-,g:65532:rwx,m::rwx,o::r-- 2022-04-13T04:59:41.926 INFO:tasks.workunit.client.0.smithi167.stderr:+ shift 2022-04-13T04:59:41.927 INFO:tasks.workunit.client.0.smithi167.stderr:++ /home/ubuntu/cephtest/mnt.0/client.0/tmp/tmp/../pjd-fstest-20090130-RC/tests/xacl/../../fstest getfacl fstest_73fed23d6d34bccdc1d73b05e3d22103/fstest_da3b235a078085ef61f2cae3419311a2 access 2022-04-13T04:59:41.927 INFO:tasks.workunit.client.0.smithi167.stderr:++ tail -1 2022-04-13T04:59:41.928 INFO:tasks.workunit.client.0.smithi167.stderr:+ r=u::r--,u:root:r-x,g::r--,g:root:rw-,m::r-x,o::r-- 2022-04-13T04:59:41.928 INFO:tasks.workunit.client.0.smithi167.stderr:+ echo u::r--,u:root:r-x,g::r--,g:root:rw-,m::r-x,o::r-- 2022-04-13T04:59:41.929 INFO:tasks.workunit.client.0.smithi167.stderr:+ egrep '^u::r--,u:root:r-x,u:65531:rwx,g::r--,g:root:rw-,g:65532:rwx,m::rwx,o::r--$' 2022-04-13T04:59:41.929 INFO:tasks.workunit.client.0.smithi167.stderr:+ '[' 1 -eq 0 ']' 2022-04-13T04:59:41.929 INFO:tasks.workunit.client.0.smithi167.stderr:+ echo 'not ok 80'
Updated by Xiubo Li almost 2 years ago
- Copied from Bug #55331: pjd failure (caused by xattr's value not consistent between auth MDS and replicate MDSes) added
Updated by Xiubo Li almost 2 years ago
This will fix the following two bugs:
https://tracker.ceph.com/issues/55240
https://tracker.ceph.com/issues/55331
Updated by Xiubo Li almost 2 years ago
Xiubo Li wrote:
This will fix the following two bugs:
https://tracker.ceph.com/issues/55240
https://tracker.ceph.com/issues/55331
The kclient has fixed them already.
Updated by Xiubo Li almost 2 years ago
- Status changed from In Progress to Fix Under Review
- Pull request ID set to 46412
- Component(FS) Client added
Updated by Venky Shankar almost 2 years ago
Xiubo Li wrote:
Xiubo Li wrote:
This will fix the following two bugs:
https://tracker.ceph.com/issues/55240
https://tracker.ceph.com/issues/55331The kclient has fixed them already.
Right - this doesn't really fix tracker #55240, isn't it?
Updated by Xiubo Li almost 2 years ago
Venky Shankar wrote:
Xiubo Li wrote:
Xiubo Li wrote:
This will fix the following two bugs:
https://tracker.ceph.com/issues/55240
https://tracker.ceph.com/issues/55331The kclient has fixed them already.
Right - this doesn't really fix tracker #55240, isn't it?
The tracker #55240 has two issues when going through the logs:
1, mds: stuck 2 seconds and keeps retrying to find ino from auth MDS as the title says.
2, Choosing random MDSes for getattr when the x caps is issued. I have fixed in kclient already in https://github.com/ceph/ceph-client/commit/5eed80fba65cd707075892450bc5d6bd464862a0, and this PR will fix it in libcephfs.
Updated by Venky Shankar almost 2 years ago
- Status changed from Fix Under Review to Pending Backport
Updated by Backport Bot almost 2 years ago
- Copied to Backport #56109: quincy: client: choose auth MDS for getxattr with the Xs caps added
Updated by Backport Bot almost 2 years ago
- Copied to Backport #56110: pacific: client: choose auth MDS for getxattr with the Xs caps added
Updated by Xiubo Li almost 2 years ago
- Status changed from Pending Backport to Resolved