Tiramisu Mokka

In the log I can see the following record ``` 08 Feb 2022 02:06:37 (pool-cta-tape-li404a) [5de80647-3a9a-4f3e-ba8a-0b2a4d7cec81 000005515735DB2942C1BA5021371B6476C5] HSM script failed (script reported: 1: sh: line 0: kill: (28254) - No...

Thanks for the info. I don't think that the QueueStat affects this, but you never now.. Can you produce a thread dump when a pool runs into such situation?

Thanks good to know. This is not quite clear to me how this change affects the flushing, But yes, might be a race or missing notification. :man_shrugging:

@ageorget can we do an experiment: start a single flush process, then cancel it with `st kill` and check that it gone with `st ls` and process is gone as...

Hi @ageorget , we are working on this. Unfortunately, any attempts to reproduce it in a lab environment have failed.

Hi @ageorget , probably you have already mentioned that, but I can't find it. Then flushes are stuck what is the result of `st ls` command and pool's `info` command....

Thanks. Looks like I got a reproduser. 🤞

@ageorget Not a 100% fix for issues at your site, but probably a partial fix of it is ready. Are you able to run one pool with a patched rpm?

@ageorget great! Can you give a try to https://sas.desy.de/index.php/s/9QjHskp7xSrEsNb

@ageorget Any update? Do you still get flushes to stuck?