-
**What happened**: When using ```shuffle = 'disk'``` merging took 50 minutes compared to 2 minutes when using ```shuffle = 'tasks'```. Also, Dask dashboard was showing very low CPU utilization when us…
-
(Reporting this as an issue mainly so that there is a canonical place it's documented, for internal reference and for posterity.)
librdkafka provides several partitioners out-of-the-box:
- `rd_kaf…
-
Clearly, on spinning disks with their very long seek times, Scylla could never achieve stunning random-access small read performance, nor very low latencies on loaded systems. However, I think we shou…
-
It took me a while to find a USB-C cable I could borrow when I needed it, and there's a bunch of random cables that I haven't been able to sort out.
For now, I've separated Lightning cables and Mic…
-
**Is your feature request related to a problem? Please describe.**
Tasks on an activityTaskqueue are randomly fetched by competing activity workers reading from the same queue.
In case of subsequent a…
-
### Background
During model initialization, each PP stage initializes their own portion of the model. This is preferrable over initializing the model all in one place and then transferring it to e…
-
### Problem Statement
There's a noticeable concern with the current method of allocating shards to nodes. Right now, there appears to be no randomness in how shards are assigned (see also example bel…
-
I know there's the [documentation ](http://docs.janusgraph.org/0.1.0-SNAPSHOT/config-ref.html) but it's unclear how it relates to `ids.block-size`. Can someone add more colour to the documentation ple…
-
Why cannot use pg_pathman?
detail error:
```
pipeline=# CREATE TABLE journal (
pipeline(# id SERIAL,
pipeline(# dt TIMESTAMP NOT NULL,
pipeline(# level INTEGER,
pipeli…
-
The challenges described in #60 arises because `context.get_person_id(usize_id)` returns `PersonId{ id: usize_id}` even if that person doesn't technically exist. Given a valid `PersonId`, we can query…