You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It seems like on dump commands, or at least rl, if the connection to the device is faster than the disk write speed, a buffer grows in RAM, and if the partition being dumped is big enough this can lead to edl getting OOM killed. For example, on my laptop with ~16 GiB combined RAM+swap, I can dump userdata to my SSD, but if I try to dump to my (slower) SD card I run out of RAM.
Naively I think it would be smart to limit the RAM usage, and pause mid-dump when needed to allow the disk cache to flush.
Couldn't find a report of this with a cursory search, but apologies if I missed one.
The text was updated successfully, but these errors were encountered:
It seems like on dump commands, or at least
rl, if the connection to the device is faster than the disk write speed, a buffer grows in RAM, and if the partition being dumped is big enough this can lead to edl getting OOM killed. For example, on my laptop with ~16 GiB combined RAM+swap, I can dump userdata to my SSD, but if I try to dump to my (slower) SD card I run out of RAM.Naively I think it would be smart to limit the RAM usage, and pause mid-dump when needed to allow the disk cache to flush.
Couldn't find a report of this with a cursory search, but apologies if I missed one.
The text was updated successfully, but these errors were encountered: