My nodeBB website is not working anymore

[wellenreiter]

@MJ
Check your redis logs or the syslog for such an entry:

Nov 11 11:26:05 web02 redis[3992]: WARNING overcommit_memory is set to 0! Background save may fail under low memory condition. To fix this issue add 'vm.overcommit_memory = 1' to /etc/sysctl.conf and then reboot or run the command 'sysctl vm.overcommit_memory=1' for this to take effect.

Basically it tells you what to do to avoid the problem.
We had the problem twice in two days, applied that setting and the forum is running fine for nearly a week now

[CreeperFace]

@HolyPhoenix said:

Also Digital Ocean works just fine, no reason to switch hosts as @CreeperFace suggests as this problem has nothing to do with your hosting provider.

I actually suggested him that because it was not expensive (free at the moment) and works like a charm. No real reason to.

[MJ]

@CreeperFace said:

m that because it was not expe

I dug into the redis security and installed a firewall with UFW, denying all incoming connections except HTTP/HTTPS/SSH and allowing all outgoing)
I also put a password on redis and bound it to 127.0.0.1
Will this be sufficient security?
Thanks for all the help, btw! Appreciate it.

[julian]

Do you have backups?

[MJ]

@julian said:

Do you have backups?

only from 4-november

[julian]

You should make regular automated backups.

[MJ]

@julian said:

You should make regular automated backups.

I know, but I have digital ocean and they support automatic backup weekly, not daily

[pichalite]

@MJ you can set up automatic backups for your redis db. Don't need backup service from digital ocean.

[MJ]

@pichalite said:

@MJ you can set up automatic backups for your redis db. Don't need backup service from digital ocean.

How do I do that?

[Bri]

@MJ Crontab is typically the way to go here

[wellenreiter]

I use logrotate to copy the database and then compress it on a daily basis.
It automatically keeps 7 days, so I cover the time from one provider backup to the next one.

[julian]

@wellenreiter You know, I never really got the hang of logrotate. Never liked the config style, but perhaps I'll look into it again.

[wellenreiter]

@julian

This is the file I use:

/var/lib/redis/*.rdb {
        rotate 1
        daily
        missingok
        copy
        maxsize 500k
}
/var/lib/redis/*.rdb.1 {
        rotate 7
        daily
        missingok
        compress
        minsize 10k
 }

It first copies the rdb file to rdb.1 and then compresses it.
7 copies of the compressed file are kept

[julian]

Does that back up the redis dump file daily, and then extend the backup to 7 days?

Is it not possible to just use a single block for it? rotate 7 on the *.rdb?

[wellenreiter]

If you are fine with just a copy you can use a rotate 7 in the first block and delete the 2nd one

[julian]

Ooh, I see... it gzips it in the 2nd block. Understood, thank you

[wellenreiter]

There is no CopyZip in logrotate, so you have to do it in two steps