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
[81X] Fix workflow 4.76 #15304
Comments
A new Issue was created by @smuzaffar Malik Shahzad Muzaffar. @davidlange6, @smuzaffar, @davidlt, @Dr15Jones can you please review it and eventually sign/assign? Thanks. cms-bot commands are list here #13029 |
On 7/27/16 4:05 AM, Malik Shahzad Muzaffar wrote:
Just add both RECO and reRECO. The logic in the setup of these workflows |
I dropped reRECO and although it fixed the step2 but then step3 failed with error message
The cmsDriver commands for step2 and step3 are
|
@wddgit David, in this case do we fail to drop the history? |
The current behavior is this. Nothing is ever dropped from the ProcessHistory, not even if all the products are dropped. In addition, it is an error to have a process read a file which has in its ProcessHistory an earlier process with the same name. The only exceptional case to mention is that if no product is produced at all, then the restriction does not apply and the process name does not get added to the ProcessHistory. I ran the code, printed out the contents of the ProcessHistory and this is the behavior I see. From looking at the code quickly, it appears this has been the behavior since at least 2011 when I put in the set of changes that implemented the reduced ProcessHistoryID. It is hard to trace back earlier, because so much of the relevant code changed at that point. I do not remember whether the same behavior was there before or not. If it was a change, then I do not remember whether it was done for a reason or was unintentional. If we wanted to change the current behavior, we would need to ensure that we dropped all the products from the process AND all subsequent processes before we tried to reuse the process name. And I think also we would need to spend some time thinking about all the metadata like ProductIDs and parentage and how they could be affected by this. It would take some time to think all this through. I can see two options in how to proceed. Option 1. We could drop the process name from the ProcessHistory, then one issue would be that events which were treated as being in different runs because they had differing ProcessHistoryIDs might suddenly have the same ProcessHistoryID because we dropped the process that caused a difference. We would have to think about how to deal with that case. Option 2. Allow the same process name to appear multiple times in the ProcessHistory. Doing this would also require a lot of thought to verify that we are not breaking something by allowing this behavior. (This is all independent of the recent bug fix I put in related to duplicate process names. The problem above related to a check run by the InputSource when reading the first run. The recent bug fix was related to a check run when initializing the lookup tables in the ProductRegistry. That was only checking entries in the input product registry that actually conflicted with the current process name. So that check would be needed in any case.) |
Why not just edit the step3 of this workflow and tell it to use |
#15498 should fix 4.76 |
assign core |
+1 |
New categories assigned: core @Dr15Jones,@smuzaffar you have been requested to review this Pull request/Issue and eventually sign? Thanks |
This issue is fully signed and ready to be closed. |
I confirm that by using
workflow 4.76 runs fine. I see that current
comes from
but updating HLTDSKIM to have reRECO means all 4.xx workflows will be dropping reRECO instead of RECO and ends up failing e.g 4.37 then fails with similar message
@davidlange6 , any idea how to avoid this?
The text was updated successfully, but these errors were encountered: