Bug #30
closedmultimds: slave_request on getattr
90%
Description
10.04.13 12:39:39.111206 408a7950 -- 10.0.1.245:6802/14233 <== client4122 10.0.1.141:0/3071614701 46048 ==== client_request(client4122:44619 getattr pAsLsXsFs #10000019a2e) ==== 114+0+0 (2100174513 0 0) 0x7f7ea7756a00
10.04.13 12:39:39.111274 408a7950 -- 10.0.1.245:6802/14233 --> mds1 10.0.1.245:6803/14253 -- slave_request(client4122:44619 authpin) -- ?+0 0x7f7e9ced0b00
10.04.13 12:39:39.111500 408a7950 -- 10.0.1.245:6802/14233 <== mds1 10.0.1.245:6803/14253 37920 ==== slave_request(client4122:44619 authpin_ack) ==== 149+0+0 (1483065163 0 0) 0x7f7ea6f0
?
Updated by Sage Weil almost 14 years ago
I think the problem is that we authpin anything we rdlock... is that really necessary?
Updated by Sage Weil over 13 years ago
In order to change this, we need a way to wake up the replica waiter when we find out the auth changed, so we can resend the REQRDLOCK. e.g., in handle_export_notify(). there is no infrastructure for that right now.
also need to verify that we resend if we were exporting/importing, and/or the migration aborted. :/
Updated by Sage Weil over 13 years ago
- Target version changed from v0.21 to v0.22
Updated by Sage Weil over 13 years ago
- Status changed from New to In Progress
- % Done changed from 0 to 50
Updated by Sage Weil over 13 years ago
- Status changed from In Progress to Resolved
commit:1d8770be2747c43a1aef05a6439b7e23a87f0419
Updated by Sage Weil over 13 years ago
Sigh.. that was a bad decision. Reverted by commit:b7b77ffc1718ea1743e199745e2c55f99ddaa16a, although if we ever go back, a few fixes are in commit:5f6d524df7a20b2d3589226669129eea04739f51.
Updated by John Spray over 7 years ago
- Project changed from Ceph to CephFS
- Category deleted (
1) - Target version deleted (
v0.22)
Bulk updating project=ceph category=mds bugs so that I can remove the MDS category from the Ceph project to avoid confusion.