Cache entries fail to be removed from Redis `cachedb` backend with `unbound-control flush* +c`

**Describe the bug**
When calling `unbound-control flush* +c` to remove cache entries from both the first-level cache and the `cachedb` second-level cache, the operation fails:

```
info: cachedb msg remove example.com A IN
debug: redis_store 9AB54C417C77A0529023C6E81D124446753F76329B01977285F8AB0D1223440E (4 bytes) with ttl 0
error: redis: set resulted in an error: ERR invalid expire time in setex
debug: redis_store set completed
```

**To reproduce**
Steps to reproduce the behavior with Redis `cachedb` backend enabled:
1. Query for a ressource record:  
`% dig @127.0.0.1 example.com A`
2. Observe Redis key being created with proper TTL in the logs:  
`debug: redis_store 9AB54C417C77A0529023C6E81D124446753F76329B01977285F8AB0D1223440E (259 bytes) with ttl 300` 
3. Flush record from both cache levels:  
`% unbound-control flush_type +c example.com A`
4. Observe the error from the logs:  
```
debug: redis_store 9AB54C417C77A0529023C6E81D124446753F76329B01977285F8AB0D1223440E (4 bytes) with ttl 0
error: redis: set resulted in an error: ERR invalid expire time in setex
```
5. Query for the same ressource record:  
`% dig @127.0.0.1 example.com A`
6. Observe the key still exists from the logs:  
```
debug: redis_lookup of 9AB54C417C77A0529023C6E81D124446753F76329B01977285F8AB0D1223440E
debug: redis_lookup found 259 bytes
```
7. Additionally, observe the key still exists in Redis with a non-null TTL:
```
% redis-cli GET 9AB54C417C77A0529023C6E81D124446753F76329B01977285F8AB0D1223440E
"\x00\x00\x81\x80\x00\x01\x00\a\x00\x00\x00\x01\aexample\x03com\x00\x00\x01\x00\x01\xc0\x0c\x00\x01\x00\x01\x00\x00\x01,\x00\x04\x17\xd7\x00\x8a\xc0\x0c\x00\x01\x00\x01\x00\x00\x01,\x00\x04`\a\x80\xaf\xc0\x0c\x00\x01\x00\x01\x00\x00\x01,\x00\x04`\a\x80\xc6\xc0\x0c\x00\x01\x00\x01\x00\x00\x01,\x00\x04\x17\xc0\xe4P\xc0\x0c\x00\x01\x00\x01\x00\x00\x01,\x00\x04\x17\xc0\xe4T\xc0\x0c\x00\x01\x00\x01\x00\x00\x01,\x00\x04\x17\xd7\x00\x88\xc0\x0c\x00.\x00\x01\x00\x00\x01,\x00_\x00\x01\r\x02\x00\x00\x01,g\xf2\x1c|g\xd6\xc2\xc5<\x88\aexample\x03com\x00\xc82{6\xff\x8c\n\xd0\x8b\xbe\x86\xcf E WG\xbc\x81\r\xff\x8fLK\xe6\x15\xc2\xf9r\x17\xbf\xad\xe0\xf8\"\xa9\xb5K\x99\xd6\xb5\x14E\x93\xdc\x96\x89\xd2\xcae\x7f\x9d\x0bO\xcd<\xa8J(\x01C\b\x14\xf7\x00\x00)\x04\xd0\x00\x00\x80\x00\x00\x00\x00\x00\x00\x00g\xddD\xdf\x00\x00\x00\x00g\xddF\x0b"
% redis-cli TTL 9AB54C417C77A0529023C6E81D124446753F76329B01977285F8AB0D1223440E
(integer) 244
```

**Expected behavior**
After executing step 3. the key(s) shall no longer exist in the Redis database.

**System:**
 - Unbound version: 1.22.0
 - OS: Debian 12
 - Redis 7.2.7 / KeyDB 6.3.4
 - `unbound -V` output:
```
Version 1.22.0

Configure line: --build=x86_64-linux-gnu --prefix=/usr --includedir=${prefix}/include --mandir=${prefix}/share/man --infodir=${prefix}/share/info --sysconfdir=/etc --localstatedir=/var --disable-option-checking --disable-silent-rules --libdir=${prefix}/lib/x86_64-linux-gnu --runstatedir=/run --disable-maintainer-mode --disable-dependency-tracking --with-pythonmodule --with-pyunbound --enable-subnet --enable-dnstap --enable-systemd --enable-cachedb --with-libhiredis --with-libnghttp2 --with-chroot-dir= --with-dnstap-socket-path=/run/dnstap.sock --disable-rpath --with-pidfile=/run/unbound.pid --with-libevent --enable-tfo-client --with-rootkey-file=/usr/share/dns/root.key --enable-tfo-server
Linked libs: libevent 2.1.12-stable (it uses epoll), OpenSSL 3.0.15 3 Sep 2024
Linked modules: dns64 python cachedb subnetcache respip validator iterator
TCP Fastopen feature available

BSD licensed, see LICENSE in source package for details.
Report bugs to unbound-bugs@nlnetlabs.nl or https://github.com/NLnetLabs/unbound/issues
```

**Additional information**

The issue comes from the fact that the following commands are refused by Redis:
```
> SETEX mykey 0 "value"
(error) ERR invalid expire time in 'setex' command
> SET mykey "value" EX 0
(error) ERR invalid expire time in 'set' command
```

However using the `EXPIRE` command works:
```
> EXPIRE mykey 0
(integer) 1
> GET mykey
(nil)
```

To fix this either:
- `1` can be passed instead of `0` in `cachedb_msg_remove_qinfo` on [`cachedb/cachedb.c:1034`](https://github.com/NLnetLabs/unbound/blob/master/cachedb/cachedb.c#L1034): key(s) will then expire in the next second
- additional logic can be added to [`cachedb/redis.c:373`](https://github.com/NLnetLabs/unbound/blob/master/cachedb/redis.c#L373) so that, when `ttl==0`, the `EXPIRE` command is invoked instead of `SET`/`SETEX`.

As the fix is specific to the Redis backend and may have less side effects, the second option is preferable.

Note that [`SETEX` is deprecated](https://redis.io/docs/latest/commands/setex/) and shall be replaced by `SET key value EX seconds`).

This issue occurred with both Redis 7.2.7 and KeyDB 6.3.4 backends for the `cachedb` module.


Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

Cache entries fail to be removed from Redis `cachedb` backend with `unbound-control flush* +c` #1253

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Uh oh!

Cache entries fail to be removed from Redis cachedb backend with unbound-control flush* +c #1253

Description

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Cache entries fail to be removed from Redis `cachedb` backend with `unbound-control flush* +c` #1253