18:31:30 GMT <barroca> hello, is it possible to go through all keys and find a particular string?
18:32:59 GMT <minus> barroca: the key names? yes, KEYS/SCAN. the values? build your own thing using KEYS/SCAN. note that it's not good to use KEYS/SCAN in production
18:33:27 GMT <barroca> @minus, no, the values itself.
18:34:37 GMT <barroca> some users of our website had complained that the were seeing other users information, just like a collision on the table.
18:35:16 GMT <barroca> but we think it is an application bug, so i need to verify how many keys have a particular value.
18:36:40 GMT <minus> yeah, you need to script something
18:36:56 GMT <Habbie> there is no 'dump everything to a text file' command either, right?
18:37:02 GMT <minus> no
18:37:16 GMT <Habbie> but again easy to build based on KEYS or SCAN
18:37:22 GMT <minus> exactly
18:37:29 GMT <barroca> is there something like for *: GET $_ ?
18:37:39 GMT <Habbie> https://github.com/p/redis-dump-load
18:37:42 GMT <Habbie> this might help?
18:37:48 GMT <minus> no, no wildcards for keys
18:37:58 GMT <minus> as that would require a full keyspace scan
18:38:04 GMT <Habbie> minus, well KEYS has wildcards, right
18:38:15 GMT <minus> and it requires a full keyspace scan
18:38:30 GMT <barroca> Habbie, i think it will work.
18:38:32 GMT <minus> KEYS and SCAN are the only commands that do that
18:38:33 GMT <Habbie> and then your client still needs to actively gets the values
18:38:37 GMT <Habbie> barroca, good
18:39:06 GMT <barroca> btw, there is no bugs on redis regarding collision right?
18:39:16 GMT <minus> i'd clone the production redis to my local machine and write a little lua script to do the search and then run that
18:39:28 GMT <Habbie> good call
18:39:42 GMT <barroca> :thumbs-up:
18:39:44 GMT <barroca> :)
18:39:46 GMT <Habbie> minus, with SLAVEOF?
18:39:52 GMT <minus> (a clone because that script can lock up redis if it takes longer)
18:39:54 GMT <minus> yes
18:39:59 GMT <Habbie> and SLAVEOF won't lock the master
18:40:07 GMT <Habbie> as it does little units of work for as long as needed
18:40:21 GMT <minus> you can slaveof no one after the dump is complete too
18:40:26 GMT <Habbie> ack
18:41:19 GMT <Habbie> minus, as you seem to know a lot, do you know if it's normal for the project to not respond to pull requests at all for a week?
18:42:04 GMT <minus> that i don't know, sorry
18:42:20 GMT <minus> but i'm used to PRs sitting waiting for years from other projects ._.
18:42:36 GMT <Habbie> sure
18:42:41 GMT <Habbie> i myself have been guilty of this as well
18:42:44 GMT <Habbie> but not at redis scale
18:42:46 GMT <Habbie> no pun intended :)
18:42:59 GMT <Habbie> i'm about to tell our local packager to just include my patch
19:00:37 GMT <barroca> dumping.. i'm using this: https://github.com/delano/redis-dump
19:03:12 GMT <minus> barroca: https://sr.ht/WLXH.txt
19:05:13 GMT <Habbie> what's .ok, minus?
19:05:44 GMT <minus> it returns a table where the value is in the ok field
19:06:00 GMT <minus> that had me puzzled for a bit
19:07:56 GMT <Habbie> huh, indeed it does
19:08:16 GMT <Habbie> which you don't notice when you just return redis.call("TYPE", ...)
19:08:22 GMT <minus> yup.
19:08:33 GMT <Habbie> also you can't return kv tables from lua it appears
19:08:35 GMT <Habbie> a pity
19:08:44 GMT <Habbie> well you can but redis will just take the values
19:08:48 GMT <Habbie> another way to not notice this
19:09:25 GMT <Habbie>      * Status replies are returned as single element table with 'ok'
19:09:27 GMT <Habbie>          * field. */
19:09:43 GMT <minus> i didn't consider TYPE to be a status reply
19:10:27 GMT <Habbie> well even then i wouldn't have known
19:10:33 GMT <Habbie> as the docs don't mention .ok at all
19:11:49 GMT <minus> a simple string reply apparently is a status reply (as opposed to the binary-safe bulk reply)
19:12:12 GMT <Habbie> was just finding that out
19:12:14 GMT <Habbie> by tcpdump
19:12:20 GMT <Habbie> looks like they conflated those things a bit
19:12:43 GMT <Habbie> and the docs on simple strings confuse non binary safe with binary safe
19:12:55 GMT <Habbie> When Redis replies with a Simple String, a client library should return to the caller a string composed of the first character after the '+' up to the end of the string, excluding the final CRLF bytes.
19:13:02 GMT <Habbie> so a distinction that a client will never learn
19:13:06 GMT <Habbie> but that suddenly matters in EVAL
19:15:22 GMT <tinyhippo> can anyone suggest what causes "Failed opening the RDB file root (in server root dir /var/spool/cron)", and how to resolve it?
19:15:33 GMT <tinyhippo> I've ran redis as a cache up to this point with no issues so far
19:15:37 GMT <minus> your redis instance got compromised
19:16:17 GMT <tinyhippo> can you provide some more detail on that, minus?
19:16:22 GMT <minus> make sure your instance is not reachable from the internet and check if your machine is infected
19:16:24 GMT <minus> hold on
19:17:16 GMT <minus> https://web.archive.org/web/20160419180101/http://antirez.com/news/96 hm, that's not quite it
19:21:06 GMT <minus> hm, not finding something better. anyway, the main issue is that your instance is reachable from the internet without password and someone changed your configuration in such a way that it tries to write a cronjob that infects your machine or so
19:21:10 GMT <minus> ^ tinyhippo
19:24:09 GMT <tinyhippo> ah
19:24:11 GMT <tinyhippo> thanks minus
19:24:15 GMT <tinyhippo> I'll sort that now
19:28:02 GMT <minus> actually that link was correct
19:28:13 GMT <minus> minus the archive.org maybe