Rohan Nagar

The source code from the dependencies is _often_ checked in to the repo as well, so you're right there is a lot of vendoring. In some cases developers choose to...

Hi @tishun could you please take a look at the linked PR and provide feedback on this performance improvement? Thank you!

Happened again today, here's some more data. Running `memory purge` does not help either. `INFO ALL` ``` # Server redis_version:7.2.5 redis_git_sha1:ba8435db redis_git_dirty:0 redis_build_id:3074ab5ecf741c27 redis_mode:standalone os:Linux 4.18.0-553.22.1.el8_10.x86_64 x86_64 arch_bits:64 monotonic_clock:POSIX clock_gettime...

Hi @sundb before the upgrade, memory would reach maxmemory and stay there. RSS memory would never climb like this, definitely never climbed much higher than the dataset memory. Our keyspace...

Hi @sundb, have you had a chance to take a look at this?

Hi @sundb was the original issue in version 7.2 ever addressed? You mentioned this: > It may be related to the change in the size (224bytes and 320 bytes) of...

Hi @sundb our defrag settings are mostly default, except for `active-defrag-cycle-max` ``` ########################### ACTIVE DEFRAGMENTATION ####################### # Enabled active defragmentation # activedefrag no activedefrag yes # Minimum amount of fragmentation...

I don't know if any of these help. Taken from a host currently experiencing high fragmentation ratio, though the system is not out of memory yet. I will get another...

Ok, let me get some better logs to share this week.

Hi @sundb thank you so much for identifying this issue and subsequent fix! I will try the upgrade to 7.4 and report back.