Hi Eric, oh, the pastebin shows how an available file looks (very small file which got uploaded as multipart). The unavailable file has the "_multipart" rados object missing, everything else is very close. I could have phrased that better. The customer now checked all their files and they are missing only three files. They've got backups and they are checking if the files are in the backup. Around the creation date of these missing files, we had problems with two OSDs which were marked down multiple times, got killed and restarted by systemd. Can this factor in the problem? We have a lot of OSD restarts, which is explained here: https://tracker.ceph.com/issues/54434 (we are now in the process of adding more and larger SSDs for the block.db devices). I am still not sure how these files could go missing. And when they exist in the backup, and their logs don't show any deletes of said files I am even more clueless. Thank you for your time and your reply. Cheers Boris Am Di., 14. Juni 2022 um 18:38 Uhr schrieb J. Eric Ivancich < ivancich@xxxxxxxxxx>: > Hi Boris, > > I’m a little confused. The pastebin seems to show that you can stat " > ff7a8b0c-07e6-463a-861b-78f0adeba8ad.81095307.17600__multipart_available_file1.pdf.2~ve8VhAEvaRSzAPfacz9rI-aLMpLY_Yw.1”, > but I thought it was missing. Can you clarify? > > The bug has been in RGW for a quite a while, well before octopus. It > involves a race condition with a very narrow window, so normally only > encountered in large, busy clusters. > > Also, I think it’s up to the s3 client whether to use multipart upload. Do > you know which s3 client the user was using? > > Eric > (he/him) > > _______________________________________________ ceph-users mailing list -- ceph-users@xxxxxxx To unsubscribe send an email to ceph-users-leave@xxxxxxx
 |