I'm a long term and experience used of POP Peeper monitoring over 20 mailboxes containing in excess of 100,000 mails.
I'm getting regular (2-3 times a week) lock-ups of the software (fails to respond, has to be forcibly crashed and re-started) and after weeks of trying to find a common denominator that actually causes the problem I think I may have found it. In simple terms I am re-sorting the entries in one mailbox at the same time as PP has started a mail retrieval run, where messages are pending deletion (i.e messages have been marked for deletion but are not actually deleted until the mail retrieval cycle) and those messages are being deleted as part of the run.
A few moments ago I tried to re-sort the messages in mailbox X (3700 messages) from date order to subject order. PP has frozen (and will have to be killed) - I now see that at the time I did this PP was in the middle of deleting messages from another mailbox as part of a retrieval run.
This last happened a couple of days ago, so I ran housekeeping on the database in case there were issues with the index, but it's just happened again.
Screenshot attached, although I am not sure it shows much.
Assistance appreciated!
Edwin
Last edited by ECourts on Fri Aug 25, 2023 4:21 am, edited 1 time in total.
ECourts -- thanks for the detailed diagnosis; it's very much appreciated!
Here's what I just tried: I added 10K messages to an account, and while PP was retrieving those messages, I would periodically toggle between which column was being sorted several times and then pause and then repeat. No problems occurred (based on your description, I wasn't necessarily expecting there to be).
Once PP had finished downloading the messages, I would mark about 10 messages at a time to be deleted, did a checkmail and then toggled the sort column between date and subject as quickly as I could. It took about 5 times, but then it happened -- PP locked up and the status bar says "deleting messages" just like you described.
CPU usage was at 0% during the lockup, which probably means that this is a deadlocking issue (e.g. 2 separate threads waiting on a resource that the other thread has already locked, so they're both waiting and cannot release the lock that the other thread needs to continue).
Again, thank you for providing the details needed to reproduce this issue. Being able to duplicate the error is crucial to fixing bugs.
Thanks Jeff, and for quickly being able to reduce the problem.
I have to be honest, this has been happening for months or maybe even years, but also fairly infrequently, but it's only recently that it became irritating enough to look further into. I think I notice it most after I have marked large numbers of messages to delete, so I guess the window of opportunity for the lock conflict is higher.
Good luck in finding a fix. if you want me to test any fix-attempts, then as a pretty advanced used, I am happy to do so. It was fun last time sorting sort
Cheers,
Edwin
I was able to find the specific problem that you discovered. As suspected, it was a deadlock issue. Then I looked for some other issues in the same area and found some more potential deadlocks (e.g. when using get next/previous unread message).
I spent the rest of the day yesterday devising and implementing a way to find these potential deadlocks semi-automatically. This method has discovered a few more potentials and some of these may just happen randomly (ie. no user interaction).
I appreciate your offer to test this, and I will take you up on that offer. However it will be at least a couple weeks before I have a build ready, so I will email you using the address you have registered on the forum (which matches email I've received from you in the past).
I had a feeling that this wasn't the only time I was getting lockups so it's good to know you found more and that hopefully you can fix them. I just never had time to do the research - It's often easier to crash it and start again, although the startup times are getting longer and longer due to the housekeeping that runs each time. I think the number of mailboxes and mails I manage are sometimes quite challenging for PP so it's good to put it through it's paces!
Looking forward to being able to test and happy to do so. Email is the same as before.
Thanks!
Edwin
The only maintenance that POP Peeper runs is compacting the databases, and that's only performed every 90 days by default (Tools / Options / Storage). The main reason why PP takes so long to load is contingent on the number of messages you have. If you want PP to load faster (and perform better overall) then moving old messages to other folders that PP doesn't check is the best solution.
