Dx create ep info updated messages #55
Conversation
|
|
||
| public function handle() | ||
| { | ||
| $this->backupStreamMessages(); |
There was a problem hiding this comment.
To be as paranoid as possible, should probably have try/catch around backing up of messages-- if this fails we wouldn't be able to restore, so shouldn't go any further.
There was a problem hiding this comment.
It occurred to me that if line 17 errors out, the following lines will not run, but in the spirit of your paranoia, I've added a check that the backup table has the same number of records as the original and throws an exception if they're not equal.
I've also wrapped the backup and check in a try/catch block to explicitly handle the case that the backup fails.
yugen
force-pushed
the
dx-create-ep-info-updated-messages
branch
from
e404657
to
1cd3657
Compare
bpow
force-pushed
the
dx-topic-replay
branch
2 times, most recently
from
598fc13
to
b6fecf0
Compare
bpow
force-pushed
the
dx-create-ep-info-updated-messages
branch
from
1cd3657
to
a49e84e
Compare
bpow
force-pushed
the
dx-create-ep-info-updated-messages
branch
from
a49e84e
to
8e08951
Compare
| public function handle() | ||
| { | ||
| // Delete errant member_retired for Aurora Pujol on group Peroxisomal Disorders VCEP | ||
| StreamMessage::find(4034)->delete(); |
There was a problem hiding this comment.
We will have to have a better way of identifying these records for deletion beside the id... because of the prior migration/reindexing, these messages will have moved to a different id (at least they appear to have done so in my test environment). Maybe they could at least be identified by timestamp?
There was a problem hiding this comment.
What if we run this in a migration before reindexing? In fact, we could run it before we do anything else to be sure.
There was a problem hiding this comment.
Lookup for these messages is now based on attributes of the message rather than the id.
…gs easier for cspec
Add actions to generate missing
ep_info_updateddx messages and reorder/reindex stream message bycreated_at.I opted to run these operations as two separate migrations so that one can fail (though that's not expected) without effecting the other.