-
Notifications
You must be signed in to change notification settings - Fork 153
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix conversations issues #5181
base: master
Are you sure you want to change the base?
Fix conversations issues #5181
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The changes make sense, but Replication.full_rename failed.
Don't know if this is an issue with the test or the actual code
This matches the logic in mailboxes.c, and makes audit tests pass
This stops us writing to the wrong DB when copying the INBOX, which fixes a long standing bug with UUID mailboxes and renames. This fixes both the imapd rename, and the replication rename codepaths.
This just makes it easier to read
@wolfsage both your comments are fixed in 75509c3 @ksmurchison - the fixes above also fixed the Replication test! Yep, looks like it was a bug before. |
int r2 = mboxlist_update_full(olddestmbentry, /*localonly*/1, /*silent*/1); | ||
if (r2) | ||
xsyslog(LOG_ERR, "IOERROR: replacing old destination tombstone after rename error", | ||
"mboxname=<%s>, error=<%s>", olddestmbentry->name, error_message(r)); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
In all four cases here should we also be logging error_message(r2) so we know why the cleanup failed?
"mboxname=<%s>, error=<%s>, cleanuperror=<%s>", oldestmbentry->name, error_message(r), error_message(r2));
There's two issues here:
basecid inconsistency causing TEMP/REAL mismatch
poor inbox detection on user rename causing bigger issues. The second one is ALSO a known Fastmail issue on user rename, such that we do a full conversations rebuild after renaming. This will fix that.
The reason these weren't detected earlier is that
ctl_conversationsdb -A
ONLY works on databases which support fetchnext, which skiplist doesn't, so we didn't notice these until switching to make twoskip the default.