I prepared a file according to redis' mass insertion protocol, and everything is working fine. Up to a point...
If the file contains up to 775 commands, all is good:
% redis-cli --pipe < in.redis
All data transferred. Waiting for the last reply...
Last reply received from server.
errors: 0, replies: 775
But starting at 776 commands, this is what I get:
% redis-cli --pipe < in.redis
Error writing to the server: No error
It seems unreasonable that there is such a tiny limit.
Plus, as you can see, the error message is fishy.
Anyone seen such a behavior? Any help would be appreciated.
I'm running the win64 version of redis (version 2.8.9)
In certain shells, there seems to be potential for interpreter errors to the \r and \n bytes that are needed for the Redis protocol. The error message you reported doesn't match exactly, but the behavior sounds somewhat similar to what is being described in this SO question:
Redis multi insert issue
Long shot, but you might try this:
echo "$(cat in.redis)" | redis-cli --pipe
Related
As I use Redis to start up with a bunch of other processes via Foreman, I find its output on startup quite verbose.
Redis writes more than twice the number of lines to stdout than any other process in my Procfile, mainly because of the ASCII art that gets printed to the log.
Is there a (startup) option to keep the log more concise, for example by turning off the output of the logo?
TLDR: If you have redis version 4.0 or higher you can do redis-server | cat to trick it into thinking it's not running in a tty.
Original answer:
I've had a quick check in the config docs and you shouldn't be seeing this. Can you maybe check your config file and see if you've set always-show-logo to yes?
The comment that accompanies it is as follows:
# By default Redis shows an ASCII art logo only when started to log to the
# standard output and if the standard output is a TTY. Basically this means
# that normally a logo is displayed only in interactive sessions.
#
# However it is possible to force the pre-4.0 behavior and always show a
# ASCII art logo in startup logs by setting the following option to yes.
I guess if you're on a version < 4.0 then that might explain what you're seeing.
Here is the issue/fix from 2014 https://github.com/antirez/redis/issues/1935
On ubuntu 16.04. Interacting with a local redis instance via redis-cli. Working with a node hubot script which uses redis as its primary data store.
when I type keys * I get a single key hubot:storage
so I FLUSHALL and get an ok response. But if the Hubot is running or else as soon as it turns on, it restores the value of that key immediately so I can never delete it.
I'v used the info command to try to see if it is persisting on some other redis instance and I've cleared all backup files from /var/redis. Basically I can't figure out where this data is being stored to keep getting restored from.
Any advice regarding how I could clear this out or where Hubot may be caching this?
It seems to be related to this code: https://github.com/hubotio/hubot-redis-brain/blob/master/src/redis-brain.js specifically the chunk at line 49 is what gets called before each restore.
Steps
Stop hubot
Flush redis (important that this is done while hubot is NOT running)
Start hubot
The reasoning is that hubot has an in-memory representation of the brain and will write it out to redis at intervals. Perhaps a nicer solution to this which would help during script development would be a command that can empty the brain and save that, but I can't see an obvious API for that in either robot.brain or hubot-redis-brain
We have a redis database running on our server, but for some reason, I cannot see any keys in our database. I'm just wondering if redis stores an activity log, where I can trace if and when the keys were deleted?
I have the usual log file for redis, at /var/log/redis.log but that doesn't have the information I am looking for.
I think there is no straight forward way to log everything but here is a hack.
$ redis-cli monitor >> ~/my_redis_commands.log 2>&1
Here >> tells OS that the output stream has been changed from monitor to a file and 2>&1 tells to redirect STDERR to STDOUT.
n>&m Merge output from stream n with stream m.
Note that file descriptor 0 is normally standard input (STDIN), 1 is standard output (STDOUT), and 2 is standard error output (STDERR).
Go and see the content of file in some SSH session for debugging.
$ tail -f ~/my_redis_commands.log
or you can use grep to find "DEL" instead. You can see the list of commands supported by Redis and try grep queries like SET, GET, etc.
$ grep '"DEL"' ~/my_redis_commands.log
Cons of this idea are:
You need to run a separate process to do this
It's memory and CPU consuming
single MONITOR client can reduce the throughput by more than 50%. Running more MONITOR clients will reduce throughput even more.
For security concerns, certain special administration commands like CONFIG are not logged into the MONITOR output
See this for more info https://redis.io/commands/monitor
The INFO command can be used to glean some forensic info when used with the all or cmdstats switch - you'll be able to see counts of all commands including offensive ones.
Keep in mind that this could be the result of an unauthorized intrusion and that your server may have been compromised.
Currently i have plan a process of using a script to add key values into redis using rpush.
I am currently using bin/sh to launch redis.
But what is the syntax to rpush my keyvalue into redis server.
#!/bin/sh
redis-cli -h 172.16.114.54
rpush stlfile "fad.stl" // how to rpush in with the correct syntax?
What kind of language is more suitable for redis so maybe i could change now to ease my future processes
Seems like using python would be easier to code Redis
from this post i found here i can easily install python.
ImportError: No module named redis
i managed to create a python script with this
import redis
r = redis.StrictRedis(host='172.16.114.54', port=6379, db=0)
r.lpush ('stlfile', 'variable.stl')
r.lrange ('stlfile', 0, -1)
when checked with another client i managed to see variable.stl went in :D
I am using Spring Jedis Client to use Redis in my application. I want to rename the commands so that no one else can fire the same just in case they are able to connect to my server.
Can anyone give an example of how to use rename command from Jedis and then how to fire subsequent commands using the the modified one ?
You can't rename a Redis command yet w/o changing the config file issue #640.
Even if you add the rename-command config file directive and restart your Redis, Jedis doesn't seem to allow sending arbitrary commands easily or to provide a trivial (i.e. no code changes) way to rename them.
What you could do, however, if you're really insistent on renaming a command and then calling it from Jedis is EVAL it. This will probably go in my pantheon of ugly hacks (:)) but after adding rename-command get foo to my /etc/redis/redis.conf and doing service redis-server restart look what I can do:
$ redis-cli
redis 127.0.0.1:6379> set bar baz
OK
redis 127.0.0.1:6379> get bar
(error) ERR unknown command 'get'
redis 127.0.0.1:6379> foo bar
"baz"
redis 127.0.0.1:6379> eval "return(redis.call('get', KEYS[1]))" 1 bar
(error) ERR Error running script (call to f_db0e060e4f58231d51f21685b20ff847de8ab9e1): Unknown Redis command called from Lua script
redis 127.0.0.1:6379> eval "return(redis.call('foo', KEYS[1]))" 1 bar
"baz"
redis 127.0.0.1:6379>
Of course, if you take this route your code can get pretty messy in no time at all so be careful where you tread... Good luck!
If a malicious user connects directly to Redis, then one can access all opcodes.
There is no feature in the Redis library to rename commands. Even if you expose access to a custom API that renames commands, you cannot change the inner opcodes of Redis itself.
Edit:
You're right, it is possible to rename commands by changing the config file indeed!
After you set the new command names, you have to recompile Jedis.
First rename the enum on src/main/java/redis/clients/jedis/Protocol.java, line 203.
Now find the corresponding enum usage on src/main/java/redis/clients/jedis/BinaryClient.java and change it also.
It may be sufficient: everywhere you still keep the old command java interfaces (e.g. zadd etc.), and inside Jedis it will talk to Redis calling the renamed command.
Is that your intention ?