[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [HTCondor-users] condor_userprio issues



Hi Ivo,

No, but we've been running 8.5.8 since December, and 8.5.6 before that, without seeing this.

I can't really upgrade this host in place. If I copy the accounting file to a host running 8.6 (or 8.7), would that be a legitimate test?

On Wed, Aug 16, 2017 at 11:37 PM, Ivo <ivo.cavalcante@xxxxxxxxx> wrote:
8.5 is a development branch, isn't? Did you try 8.4 or 8.6?


On qua, 16 de ago de 2017 18:35 Jon Bernard <jonbernard@xxxxxxxxx> wrote:
Hi Ivo,

jones is active, but this also occurs for users who are not. We're not using HA.

BTW, this is 8.5.8.

On Wed, Aug 16, 2017 at 4:20 PM, Ivo <ivo.cavalcante@xxxxxxxxx> wrote:

Jon,

Is user Jones actively using resources between dumps? Is it possible for him not to use it for some hours? What I think might be happening:

For a given user, priorities "naturally" decreases over time. This is part of balance algorithm, and can be somewhat tuned. So, ups and downs, on this case, might not be related to start date.

Nonetheless, it seems strange for condor_userprio to be changing start dates like this. Are you using high availability?

Disclaimer: never had this problem, just curious about it. So, what I said may make no sense at all, "use at your own risk". :-)


On qua, 16 de ago de 2017 16:43 Jon Bernard <jonbernard@xxxxxxxxx> wrote:
Hi all,

We've been having some issues with condor_userprio for the past couple of months.

On the first of every month, we do "condor_userprio -resetall", and thereafter we dump the usage every hour with "condor_userprio -all --allusers". Here's the usage for user jones for today:

jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 57485.04ÂÂÂ 57.49ÂÂ 1000.00ÂÂÂÂÂ 0ÂÂÂÂ 43029.67Â 7/01/2017 00:05Â 8/15/2017 13:23ÂÂÂ 0+10:37
jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 55843.64ÂÂÂ 55.84ÂÂ 1000.00ÂÂÂÂÂ 0ÂÂÂÂ 43029.67Â 7/01/2017 00:05Â 8/15/2017 13:23ÂÂÂ 0+11:37
jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 54019.48ÂÂÂ 54.02ÂÂ 1000.00ÂÂÂÂÂ 0ÂÂÂÂ 11043.93Â 8/01/2017 00:05Â 8/15/2017 13:23ÂÂÂ 0+12:37
jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 52709.88ÂÂÂ 52.71ÂÂ 1000.00ÂÂÂÂÂ 0ÂÂÂÂ 43029.67Â 7/01/2017 00:05Â 8/15/2017 13:23ÂÂÂ 0+13:37
jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 51214.68ÂÂÂ 51.21ÂÂ 1000.00ÂÂÂÂÂ 0ÂÂÂÂ 43029.67Â 7/01/2017 00:05Â 8/15/2017 13:23ÂÂÂ 0+14:37
jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 50509.25ÂÂÂ 50.51ÂÂ 1000.00ÂÂÂÂÂ 0ÂÂÂÂ 43056.22Â 7/01/2017 00:05Â 8/16/2017 04:42ÂÂÂ 0+00:18
jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 61868.16ÂÂÂ 61.87ÂÂ 1000.00ÂÂÂ 311ÂÂÂÂ 43506.80Â 7/01/2017 00:05Â 8/16/2017 06:01ÂÂÂÂÂ <now>
jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 74270.75ÂÂÂ 74.27ÂÂ 1000.00ÂÂ 1501ÂÂÂÂ 44002.47Â 7/01/2017 00:05Â 8/16/2017 07:01ÂÂÂÂÂ <now>
jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 74930.58ÂÂÂ 74.93ÂÂ 1000.00ÂÂÂÂ 62ÂÂÂÂ 12121.67Â 8/01/2017 00:05Â 8/16/2017 08:01ÂÂÂÂÂ <now>
jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 76500.23ÂÂÂ 76.50ÂÂ 1000.00ÂÂÂÂÂ 0ÂÂÂÂ 12251.81Â 8/01/2017 00:05Â 8/16/2017 08:50ÂÂÂ 0+00:10
jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 78836.92ÂÂÂ 78.84ÂÂ 1000.00ÂÂÂÂ 34ÂÂÂÂ 44389.08Â 7/01/2017 00:05Â 8/16/2017 10:00ÂÂÂÂÂ <now>
jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 76861.32ÂÂÂ 76.86ÂÂ 1000.00ÂÂÂÂ 62ÂÂÂÂ 44398.50Â 7/01/2017 00:05Â 8/16/2017 11:00ÂÂÂÂÂ <now>
jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 83441.17ÂÂÂ 83.44ÂÂ 1000.00ÂÂÂ 212ÂÂÂÂ 12724.79Â 8/01/2017 00:05Â 8/16/2017 12:01ÂÂÂÂÂ <now>
jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 87386.80ÂÂÂ 87.39ÂÂ 1000.00ÂÂÂ 218ÂÂÂÂ 44925.79Â 7/01/2017 00:05Â 8/16/2017 13:01ÂÂÂÂÂ <now>
jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 85440.72ÂÂÂ 85.44ÂÂ 1000.00ÂÂÂÂÂ 0ÂÂÂÂ 44944.94Â 7/01/2017 00:05Â 8/16/2017 13:28ÂÂÂ 0+00:33
jones@xxxxxxxÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ 82761.34ÂÂÂ 82.76ÂÂ 1000.00ÂÂÂÂÂ 0ÂÂÂÂ 12955.78Â 8/01/2017 00:05Â 8/16/2017 13:28ÂÂÂ 0+01:32

As you can see, the total usage cycles up and down, depending on the start date.

Is there a simple way to fix this for all users?

I've tried condor_userprio -setbegin $(date -d $(date '+%Y-%m-01') '+%s'), and that seems to eventually set the start date for the user to 8/01/2017 (although sometimes I need to do it multiple times before it has an effect). Is that the way to fix this? Or does it not do what I think it does (i.e., cause condor_userprio to only count usage since the start date)?

Thanks,
Jon
_______________________________________________
HTCondor-users mailing list
To unsubscribe, send a message to htcondor-users-request@cs.wisc.edu with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/htcondor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/htcondor-users/

_______________________________________________
HTCondor-users mailing list
To unsubscribe, send a message to htcondor-users-request@cs.wisc.edu with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/htcondor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/htcondor-users/

_______________________________________________
HTCondor-users mailing list
To unsubscribe, send a message to htcondor-users-request@cs.wisc.edu with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/htcondor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/htcondor-users/

_______________________________________________
HTCondor-users mailing list
To unsubscribe, send a message to htcondor-users-request@cs.wisc.edu with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/htcondor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/htcondor-users/