Hello, I am running a connection-heavy server (1,700 mailboxes, ~1,200 activesync connections) on one machine. This has worked really well as it is pretty low volume (mostly for inter-user communication). I have continued to tweak various performance metrics to tweak out whatever I can. However, recently, a new error has started showing up that I can't place: docker logs -n 100 mailcowdockerized_sogo-mailcow_1 | grep UNKNOWN Apr 9 14:10:55 35b46a50e7ee sogod [6765]: an error occurred when caching value for key '+attributes': "UNKNOWN READ FAILURE" Apr 9 14:10:55 35b46a50e7ee sogod [6765]: an error occurred when caching value for key '+attributes': "UNKNOWN READ FAILURE" Apr 9 14:10:55 35b46a50e7ee sogod [6765]: an error occurred when caching value for key +defaults': "UNKNOWN READ FAILURE" Apr 9 14:10:55 35b46a50e7ee sogod [6765]: an error occurred when caching value for key '+settings': "UNKNOWN READ FAILURE" Apr 9 14:10:55 35b46a50e7ee sogod [6765]: an error occurred when caching value for key '+attributes': "UNKNOWN READ FAILURE" Apr 9 14:10:55 35b46a50e7ee sogod [6765]: an error occurred when caching value for key '+attributes': "UNKNOWN READ FAILURE" Apr 9 14:10:55 35b46a50e7ee sogod [6765]: an error occurred when caching value for key '+defaults': "UNKNOWN READ FAILURE" Apr 9 14:10:55 35b46a50e7ee sogod [6765]: an error occurred when caching value for key '+settings': "UNKNOWN READ FAILURE" When I was looking into this last night, I noticed that I actually had two memcached servers - one from a previous version I hadn't noticed! So I ran the instructions found here: https://github.com/mailcow/mailcow-dockerized/issues/2182 docker-compose down docker rm -f $(docker ps -a -q) docker-compose up -d This removed the 5 or so old "mailcow-dockerized" instances, and I thought it might have solved this problem, but no go. I have googled this error, and I'm only finding "UNKNOWN READ FAILURE" when talking about memcached. There isn't much I can see to tweak here. Currently, this is an AWS instance running with 32GB of RAM + a 32GB swap, and 4 CPU. Like I said, things are very good with performance, just concerned about this new error popping up every few seconds.

Just an update - I think I figured this out. Seems like memcached was maxed out! Seems like a capacity issue on my server. In case it helps anyone, here is what I ended up doing: 1. confirm current settings, and any issues: ``` sudo docker-compose exec -u root memcached-mailcow sh echo stats | nc localhost 11211 echo stats | nc localhost 11211 | grep connections ``` 2. Increase number of connections (Nervous about this one - I have a feeling it should be in the override file, but haven't played enough with that yet to know...) by adding "command" line shown below: ``` memcached-mailcow: image: memcached:alpine restart: always environment: - TZ=${TZ} command: -c 2048 #this is the added line networks: mailcow-network: aliases: - memcached ``` I have not tested if my problem is gone yet (need some up time), but this feels right :)

"UNKNOWN READ FAILURE" - Memcached issue?

n1sni

Hello,
I am running a connection-heavy server (1,700 mailboxes, ₁,200 activesync connections) on one machine. This has worked really well as it is pretty low volume (mostly for inter-user communication). I have continued to tweak various performance metrics to tweak out whatever I can. However, recently, a new error has started showing up that I can’t place:

docker logs -n 100 mailcowdockerized_sogo-mailcow_1 | grep UNKNOWN
Apr 9 14:10:55 35b46a50e7ee sogod [6765]: <0×0×5629cec9b950[SOGoCache]> an error occurred when caching value for key ‘<removed>+attributes’: “UNKNOWN READ FAILURE”
Apr 9 14:10:55 35b46a50e7ee sogod [6765]: <0×0×5629cec9b950[SOGoCache]> an error occurred when caching value for key ‘<removed>+attributes’: “UNKNOWN READ FAILURE”
Apr 9 14:10:55 35b46a50e7ee sogod [6765]: <0×0×5629cec9b950[SOGoCache]> an error occurred when caching value for key <removed>+defaults’: “UNKNOWN READ FAILURE”
Apr 9 14:10:55 35b46a50e7ee sogod [6765]: <0×0×5629cec9b950[SOGoCache]> an error occurred when caching value for key ‘<removed>+settings’: “UNKNOWN READ FAILURE”
Apr 9 14:10:55 35b46a50e7ee sogod [6765]: <0×0×5629cec9b950[SOGoCache]> an error occurred when caching value for key ‘<removed>+attributes’: “UNKNOWN READ FAILURE”
Apr 9 14:10:55 35b46a50e7ee sogod [6765]: <0×0×5629cec9b950[SOGoCache]> an error occurred when caching value for key ‘<removed>+attributes’: “UNKNOWN READ FAILURE”
Apr 9 14:10:55 35b46a50e7ee sogod [6765]: <0×0×5629cec9b950[SOGoCache]> an error occurred when caching value for key ‘<removed>+defaults’: “UNKNOWN READ FAILURE”
Apr 9 14:10:55 35b46a50e7ee sogod [6765]: <0×0×5629cec9b950[SOGoCache]> an error occurred when caching value for key ‘<removed>+settings’: “UNKNOWN READ FAILURE”

When I was looking into this last night, I noticed that I actually had two memcached servers - one from a previous version I hadn’t noticed! So I ran the instructions found here: mailcow/mailcow-dockerized2182

docker-compose down
docker rm -f $(docker ps -a -q)
docker-compose up -d

This removed the 5 or so old “mailcow-dockerized” instances, and I thought it might have solved this problem, but no go.

I have googled this error, and I’m only finding “UNKNOWN READ FAILURE” when talking about memcached. There isn’t much I can see to tweak here. Currently, this is an AWS instance running with 32GB of RAM + a 32GB swap, and 4 CPU. Like I said, things are very good with performance, just concerned about this new error popping up every few seconds.

n1sni

Just an update - I think I figured this out. Seems like memcached was maxed out!

Seems like a capacity issue on my server. In case it helps anyone, here is what I ended up doing:

confirm current settings, and any issues:

sudo docker-compose exec -u root memcached-mailcow sh
echo stats | nc localhost 11211
echo stats | nc localhost 11211 | grep connections

Increase number of connections (Nervous about this one - I have a feeling it should be in the override file, but haven’t played enough with that yet to know…) by adding “command” line shown below:

memcached-mailcow:
      image: memcached:alpine
      restart: always
      environment:
        - TZ=${TZ}
      command: -c 2048  #this is the added line
      networks:
        mailcow-network:
          aliases:
            - memcached

I have not tested if my problem is gone yet (need some up time), but this feels right 🙂