Redis 4.0.14 on Windows Server 2016

[992582781]

The redis 4.0.14 sentinel is deployed on the windows 2016 server and will be down in a few hours。

[tporadowski]

@992582781 what is the issue here?

[992582781]

@992582781 what is the issue here?

Three Sentinels were running on three computers, and a day later, two of them prompted the other to go offline (sdown).

[tporadowski]

Could you please provide some more information? What is your configuration, is there anything in log files related to this behavior (either Redis' log files or in Event Viewer), any steps to reproduce...?

[992582781]

Could you please provide some more information? What is your configuration, is there anything in log files related to this behavior (either Redis' log files or in Event Viewer), any steps to reproduce...?

Sentinel configuration: port 63791

master

protected-mode no bind 192.168.100.10 sentinel monitor master 192.168.100.10 6379 2 sentinel down-after-milliseconds master 5000 sentinel failover-timeout master 10000 sentinel parallel-syncs master 1 logfile "redis_sentinel_log.txt" dir "D:\redis"

Now I use Windows Server 2012, sentinel has no problem

[tporadowski]

Recently I've released a hotfix related to running Redis for Windows in Sentinel mode - 4.0.14.2, are you using that latest version?

[992582781]

Recently I've released a hotfix related to running Redis for Windows in Sentinel mode - 4.0.14.2, are you using that latest version?

Hi，I download and install it on this link (https://github.com/tporadowski/redis/releases/tag/v4.0.14.2).

[tporadowski]

Ok, so you're using the latest. What about the logs? Was there anything specific related to to shutting down of that sentinel process?

[992582781]

Ok, so you're using the latest. What about the logs? Was there anything specific related to to shutting down of that sentinel process?

Three Sentinels were running on three computers A day later, two of them prompted the other to go offline (sdown,but sentinel is normal).

[tporadowski]

I would like to help here and resolve the issue, but I would need a way to reproduce this in the first place. If you happen to have more details or observations - please share them here.

[andrpich]

i just set up a test env running Running 4.0.14.2 on windows 2016. currently 3 nodes that will be expanded to 5 within the next days.

will keep you updated and share logs if there are any incidents.

[992582781]

Thank you. My problem was that the Sentinels couldn't switch master and slave.

[andrpich]

Thank you. My problem was that the Sentinels couldn't switch master and slave.

That is something i can reproduce:

When i manually disable the master all Sentinels do notice. - # +sdown master .

Unfortunately thats it. there is no odown logged on any node and no election of a new master.

[andrpich]

solved, but not sure if it is a configuration issue or an actual problem to deal with.

as soon as requirepass is activated in sentinel.conf a sentinel node marks all other sentinels as sdown at startup. i suppose the nodes are not able to talk to each other and therefore i did not see an odown or a voting process.

if requirepass is diabled everything works like a charm.

what i noticed:

sentinel config rewrite does affect the requirepass parameter

requirepass MYPASS is changed to requirepass "MYPASS" in some but not all nodes.

[andrpich]

Update:

still running smoothly. im am able to switch back and forth master without any issues.

i will have a closer look into the requirepass thing tomorrow.

[992582781]

OK，your reply, I have read it, thank you

[andrpich]

no further findings. i suppose this can be closed.