Install instructions give error at step to install redis on mac

[AmandaBirmingham]

The install instructions at

qiita/INSTALL.md

Line 119 in 4492c3d

brew install homebrew/versions/redis28

give the command
brew install homebrew/versions/redis28
but Homebrew 4.1.4 (current version as of 08/11/2023) gives the error:
Error: homebrew/versions was deprecated. This tap is now empty and all its contents were either deleted or migrated.

@charles-cowart has indicated that it should be ok to use the latest version of redis instead (current up around 7.0), in which case the command in the install could change to:
brew install redis

[charles-cowart]

I think @antgonza should weigh in.

[wasade]

@charles-cowart, could you verify the redbiom test suite works with 7.0? It would be great if that is the case. On production I think we're using 3.2.6

@AmandaBirmingham, redis is also available from conda-forge

[AmandaBirmingham]

These may be the fastest issue responses I've ever seen :) :)

[antgonza]

FWIW, on my computer I'm using:

$ redis-server --version
Redis server v=5.0.3 sha=00000000:0 malloc=libc bits=64 build=499afec3908d6e99

and everything is working fine. Additionally, I didn't find in my email or notes any reason to keep using that older version.

I'm gonna issue a PR using the latest version and see what happens.

[antgonza]

Never mind, there are no restrictions in Qiita and the builds have been using higher version for both qiita and redbiom.

[antgonza]

Finally getting back to this.

Is it possible to let homebrew install whatever the newest version of redis is? If yes, would everyone agree that this could be a good solution for this issue? If not, what could it be a good solution?

[AmandaBirmingham]

Works for me :)

[charles-cowart]

Finally getting back to this.

Is it possible to let homebrew install whatever the newest version of redis is? If yes, would everyone agree that this could be a good solution for this issue? If not, what could it be a good solution?

Agreed, it works for me as well. I believe 'brew install redis' will install the latest by default.

[wasade]

It assumes people use brew though which I think introduces a new developer expectation. Is there a reason the conda-forge build of redis doesn't work?

[wasade]

oh weird, brew is already part of the installation instructions? Is that redundant with conda?

[antgonza]

Honestly, I think is historical ... to when things were hard to install in M1/M2 ...

[charles-cowart]

@wasade it is also installable through conda (conda install redis), docker containers, and perhaps even other avenues. As long as the latest version can reliably be used then I don't think it matters so much how it's installed. Pinning the version to 2.x is what it makes it harder to install as it becomes unsupported on most of these installation pathways.

[wasade]

Yes, but we currently assume conda on deploy whereas we do not assume docker or brew (AFAIK).

Pins are hard and should be removed unless needed

[charles-cowart]

Yes, but we currently assume conda on deploy whereas we do not assume docker or brew (AFAIK).

Pins are hard and should be removed unless needed

It sounds like we're all in agreement that this one can be removed then.

[wasade]

Possibly, were you able to perform the check discussed here?

[charles-cowart]

Confirmed the tests do pass using Redis 5.0.3 from the conda environment and using Redis 7.2.4 installed with homebrew. The number of messages sent using pubsub was changed to 1000 to shorten the test time; all other defaults were used. Not sure why two tests were skipped - do you know?

5.0.3:
(redbiom_test) ccowart@Applejack tests % python basic.py
..................ss.....

Ran 25 tests in 0.034s

OK (skipped=2)
(redbiom_test) ccowart@Applejack tests % bash bench.sh
Testing on 127.0.0.1:7379 with 100 clients in parallel, for a total of 100,000 requests per benchmark.
PING: 58356.06 requests/sec.
SET(hello,world): 55115.43 requests/sec.
GET(hello): 56811.69 requests/sec.
INCR(hello): 57226.83 requests/sec.
LPUSH(hello,abc): 57086.33 requests/sec.
LRANGE(hello,0,10): 57221.14 requests/sec.
LRANGE(hello,0,100): 48426.13 requests/sec.
(redbiom_test) ccowart@Applejack tests % ./pubsub -n 1000
Received 1008 messages from 10 writers to 450 readers through 1 channels in 0.00 sec: received inf msg/sec
(redbiom_test) ccowart@Applejack tests % ./websocket
After 0.06 sec: 3,329 messages sent, 10,000 received (100.00%). Instant rate: 169,765/sec, overall rate: 169,765/sec
(redbiom_test) ccowart@Applejack tests % bash curl-tests.sh
Test 1/2: Large PUT upload: generating the key... OK
Sending a PUT request with a large payload... OK
Verifying the upload... OK

Test 2/2: Sending an OPTIONS request... OK
(redbiom_test) ccowart@Applejack tests % python limits.py
test_huge_upload (main.TestHugeUrl) ... ok
test_huge_url (main.TestHugeUrl) ... ok

---

Ran 2 tests in 0.204s

OK
(redbiom_test) ccowart@Applejack tests %

7.2.4:
(redbiom_test) ccowart@Applejack tests % python basic.py
..................ss.....

Ran 25 tests in 0.030s

OK (skipped=2)
(redbiom_test) ccowart@Applejack tests % bash bench.sh
Testing on 127.0.0.1:7379 with 100 clients in parallel, for a total of 100,000 requests per benchmark.
PING: 53388.40 requests/sec.
SET(hello,world): 53901.44 requests/sec.
GET(hello): 51155.26 requests/sec.
INCR(hello): 53898.65 requests/sec.
LPUSH(hello,abc): 55084.16 requests/sec.
LRANGE(hello,0,10): 54021.09 requests/sec.
LRANGE(hello,0,100): 43617.36 requests/sec.
(redbiom_test) ccowart@Applejack tests % ./pubsub -n 1000
Received 1008 messages from 10 writers to 450 readers through 1 channels in 0.00 sec: received inf msg/sec
(redbiom_test) ccowart@Applejack tests % ./websocket
After 0.05 sec: 3,329 messages sent, 10,000 received (100.00%). Instant rate: 187,385/sec, overall rate: 187,385/sec
(redbiom_test) ccowart@Applejack tests % bash curl-tests.sh
Test 1/2: Large PUT upload: generating the key... OK
Sending a PUT request with a large payload... OK
Verifying the upload... OK

Test 2/2: Sending an OPTIONS request... OK
(redbiom_test) ccowart@Applejack tests % python limits.py
test_huge_upload (main.TestHugeUrl) ... ok
test_huge_url (main.TestHugeUrl) ... ok

---

Ran 2 tests in 0.207s

OK

[wasade]

I think those are the redis tests, the earlier question was w.r.t. the redbiom test suite though?

[charles-cowart]

I think those are the redis tests, the earlier question was w.r.t. the redbiom test suite though?

doh! You're right, they were the webdis tests. Wrong directory.

[wasade]

No worries, thanks!

…

On Jan 9, 2024, at 12:24 PM, Charles Cowart ***@***.***> wrote: I think those are the redis tests, the earlier question was w.r.t. the redbiom test suite though? doh! You're right, they were the webdis tests. Wrong directory. — Reply to this email directly, view it on GitHub <https://urldefense.com/v3/__https://github.com/qiita-spots/qiita/issues/3310*issuecomment-1883729217__;Iw!!Mih3wA!AhgFgQAUPWNNnnF_XvhfgA_BnPixxC-Uj8q5HYgJORrJ02gfzNPOF8tTdxXoBrE15CIpfhhccYkPWZL1GjrwDRomRJ3X31yaoXKpow$>, or unsubscribe <https://urldefense.com/v3/__https://github.com/notifications/unsubscribe-auth/AADTZMWTBRZALGWTBNF2ZEDYNWRRRAVCNFSM6AAAAAA3NFJG7GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQOBTG4ZDSMRRG4__;!!Mih3wA!AhgFgQAUPWNNnnF_XvhfgA_BnPixxC-Uj8q5HYgJORrJ02gfzNPOF8tTdxXoBrE15CIpfhhccYkPWZL1GjrwDRomRJ3X31wfUNYp6Q$>. You are receiving this because you were mentioned.

[charles-cowart]

test.log
Confirmed redbiom tests completed successfully using Redis 7.2.4

[wasade]

Great, thank you!