Skip to Content.
Sympa Menu

perfsonar-user - Re: [perfsonar-user] [Monipe-des] Use of SWAP in Central Management

Subject: perfSONAR User Q&A and Other Discussion

List archive

Re: [perfsonar-user] [Monipe-des] Use of SWAP in Central Management


Chronological Thread 
  • From: Andrew Lake <>
  • To: Marcos Felipe Schwarz <>, Comunicação com os desenvolvedores do projeto MONIPE <>
  • Cc: perfsonar-user <>
  • Subject: Re: [perfsonar-user] [Monipe-des] Use of SWAP in Central Management
  • Date: Wed, 7 Aug 2019 10:45:20 -0700

Hi Marcos,

There is no reason for two to be running, it just likely means one started before the other one finished or similar.You may also want to consider increasing the “expire” values in /etc/perfsonar/toolkit/clean_esmond_db.conf so it is not dealing with as much data at once. The existence of the cleaner script at all is less than ideal, but that is a topic for another day...

We will incorporate your change prior to the 4.2.0 release since I think you rightly point out that we should be preventing two instances of those from running at the same time. Thanks!

Andy


On August 6, 2019 at 5:43:01 PM, Marcos Felipe Schwarz () wrote:

Taking a better look at the problem I noticed that each instace of ps_remove_data.py is consuming around 10G of RAM. So two instances of this script is too much for our 32GB central-management host.
1. Is there any reason for two instances of this script running at the same time?
2. Is it normal for ps_remove_data.py consume this much of RAM on the first place?

To bypass the problem I temporarely edited /etc/cron.d/cron-clean_esmond_db to:
30 2 * * * root /usr/bin/flock -n /tmp/clean_esmond_db.lockfile /usr/lib/perfsonar/scripts/clean_esmond_db.sh &> /var/log/perfsonar/clean_esmond_db.log

Can a fix like this be incorporated?

Regards,

Marcos Schwarz


From: "Murilo Vetter" <>
To: "perfsonar-user" <>
Cc: "Comunicação com os desenvolvedores do projeto MONIPE" <>
Sent: Tuesday, July 16, 2019 2:44:50 PM
Subject: [Monipe-des] Use of SWAP in Central Management

Hi list,

We are experiencing some degradation in out central management host.

It is using SWAP, as it can be observed bellow:

# free -h
              total        used        free      shared  buff/cache   available
Mem:            31G         29G        510M        144M        902M        679M
Swap:          3.7G        3.5G        264M

As we can check bellow, there is a process ps_remove_data.py that takes too long to finished and it runs again even if it has already another process running. 


Is there a way to minimize this problem? A Locking control to not run twice or another way to solve this problem?

Regards,
Murilo Vetter

_______________________________________________
Monipe-des mailing list

http://listas.rnp.br/mailman/listinfo/monipe-des
--
To unsubscribe from this list: https://lists.internet2.edu/sympa/signoff/perfsonar-user



Archive powered by MHonArc 2.6.19.

Top of Page