Nicolae Vartolomei

Fix timequery returning wrong offset after trim-prefix which could lead to stuck consumers

7

The fix works only if compression is not in use. We need follow-up work which would decompress the batches to find the exact offset to return, or (!) we could...

Fix timequery not returning results when racing with archival retention and gc

7

The meat of the PR: Tiered Storage physically has a superset of the addressable data. This can be caused at least by the following: a) trim-prefix, b) retention being applied...

Currently a segfault is reported as: ``` Segmentation fault on shard 0. Backtrace: ... ``` I believe we could add additional data about the context in which the segfault happens...

Proposal to change the semantics of the last stable offset to always mean committed offset rather than "raft last visible index"

2

In the write caching design we made the assumption that last stable offset is always equal-or-greater than the committed offset because when transactions are in use, we always wait for...

model: make kafka::next_offset saturate rather than overflow

1

This is similar to change introduced in 8c3777bb15df27a49d086e28dead63f5278946f7 but this time for kafka::offset. Quoting from the previous commit message: model::offset::max() is often used to indicate "no upper bound" on operations....

High lowres clock drift > 1s in debug mode

4

In redpanda we have a [test](https://github.com/redpanda-data/redpanda/issues/13275) which began failing very often in debug mode since latest seastar upgrade (CPU is busy with other processes too) The test gets the current...

In Seastar, metrics can be defined with implicit aggregation by specific labels, which occurs at query time. This feature is useful, for instance to define metrics per shard or even...

p command line prints invisible/white text

2

See attached screenshot ^ There is more text, it is white on white.

Liveness specification fails if a monitor is in a hot state at the end of a schedule. I have some monitors which contain state about send requests and handled ones....

net::base_transport should try multiple resolved addresses / multiple AF_INET families

2

In https://github.com/redpanda-data/redpanda/pull/18412 I have tried to disable reconnections in tight loop in the http client. Some failing tests did uncover a problem with net::base_transport. That is, it does a DNS...