OpenZFS 2.4.1rc4 Sequioa Intel Core i7 15.7.4 Send/Receive between pools panics #67
Replies: 5 comments 5 replies
|
Pool has nested datasets if that is relevant. I have not "upgraded" the pools with new 2.4.1 features (still on 2.3.1). |
|
Eeek! I downgraded to 2.3.1 and still got a panic. Maybe this isn't related to my upgrade to 2.4.1? Hmm... panic is a bit different too: |
|
which is a ZFS panic, not a macOS one. It is concerning though. You can tell it to not panic in this case, for best-effort-recover of the data though. |
|
Yeah, I was just looking around and came across this: openzfs#17307 I have the same failure on send to a different pool (I have a backup of the backups). On the second pool though, the send failure completely toasted the pool (unable to import). I've recreated both target pools and am about to retry some sends. Its distrubing that this has appeared on both destination pools at "the same time" which makes me suspicious about the source pool, however, that seems fine (scrubbed ok). Will continue to futz about. |
I rolled back to 2.3.1 (not 2.3.1p1) and memory issue seems to be gone. I've copied a ~1.1T sparse bundle from the ZVOL mentioned above to another (mirrored) pool and wired memory has stayed around 19Gb where as it would have hit 31Gb+ (and the box likely panic/crashed) before completing while running 2.4.1rc4. |
Uh oh!
There was an error while loading. Please reload this page.
Sending between to local pools panics hard (2 out of 2 attempts):
All reactions