A ton of 'yum update rockstor' processes

I noticed that my Rockstor installation was less responsive than usual this morning and found that there were a whole lot of “update rockstor” processes running and active. I assume that these have been incrementing over time and if the count goes up I will update this thread later.

[root@san01 ~]# ps aux | grep 'update rockstor' | wc -l

I’m inlined to reboot or start killing processes but it’s not so bad right now that I can’t work around it. CPU idle is still high and load averages are low around 0.25. It’s a 4 core system so I won’t hit the panic button yet. If there’s any live troubleshooting, etc. that you’d like me to do let me know.

Confirmed it is indeed incrementing.

[root@san01 ~]# ps aux | grep 'update rockstor' | wc -l

Also here is the actual process line which I should have included in the first post!

[root@san01 ~]# ps aux | grep 'update rockstor'
root       340  0.0  0.1 347852 30312 ?        S    07:58   0:06 /usr/bin/python /usr/bin/yum update rockstor --changelog --assumeno
root       513  0.1  0.1 347852 30116 ?        S    08:03   0:12 /usr/bin/python /usr/bin/yum update rockstor --changelog --assumeno
root       657  0.0  0.1 347852 30288 ?        S    08:08   0:07 /usr/bin/python /usr/bin/yum update rockstor --changelog --assumeno
root       794  0.1  0.1 347852 30376 ?        S    08:13   0:10 /usr/bin/python /usr/bin/yum update rockstor --changelog --assumeno

etc. etc.

Thats ugly. Thanks for bringing it up. Can you paste the entire process line? I am curious to see all the parameters passed to yum and the parent process etc… more info the better.

There are two types of yum processes on Rockstor. One is an update check that happens in the background when you are logged into the ui. This is not a repeat one, just happens once per your ui visit/refresh. The other is via yum cron, perhaps we need to tweek it so it doesn’t ping too often.

Thanks for the info. I think this is what’s happening. If you are logged into the webui, it issues a call to the backend every 5 minutes to see if there is an update available. Normally, the backend call returns within a few seconds. But in your case, may due to a network issue, the backend call(hence the yup update process) never returns and the frontend gives up and tries again after 5 more minutes. So eventually, you have the pile up.

Is your Rockstor box setup to talk to the internet? You can safely kill all those processes btw. No need to reboot the box.

There are a few things going on here and they will be enhanced by this

Yup the issue was that rockstor was uanble to route to the internet. I resolved that, cleared the processes and now it’s back to normal.