Multi-Site sync error with multipart objects: Resource deadlock avoided

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi,

We've been trying to set up multi-site sync on two test VMs before rolling things out on actual production hardware. Both are running Ceph 18.2.4 deployed via cephadm. Host OS is Debian 12, container runtime is podman (switched from Debian 11 and docker.io, same error there). There is only one RGW daemon on each site. Ceph config is pretty much defaults. One thing I did change was setting rgw_relaxed_region_enforcement to true because the zonegroup got renamed from "default" during the switch to multi-site using the dashboard's assistant. There's nothing special like server-side encryption either. Our end goal is to replicate all RGW data from our current cluster to a new one.

The Multi-Site configuration itself went pretty smoothly through the dashboard and pre-existing data started syncing right away. Unfortunately, not all objects made it. To be precise, none of the larger objects over the multipart threshold got synced. This is consistent for newly uploaded multipart objects as well. Curiously, it's working fine in the other direction, i.e. multipart uploads from the secondary zone do get synced to the master.

Here are some relevant logs:


[Index of Archives]     [Information on CEPH]     [Linux Filesystem Development]     [Ceph Development]     [Ceph Large]     [Ceph Dev]     [Linux USB Development]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [xfs]


  Powered by Linux