How can we address disk space issue on Unix server?

Posted on 2014-01-31
Medium Priority
Last Modified: 2014-02-01
My Unix team have just received a critical alert for one of our live Production Unix servers: jdeapp1 for swap. When they investigated they found there were few files which are growing very quickly and consuming capacity.

The files are all in /tmp folder

this is the files they provided. I'm not a Unix expert so I dont know how to proceed to help them. I have never seen these file types before.

Can anyone give me some guidance on what these files might be or how we can address this?
-rw-r--r--   1 JDE      JDE            0 Jan 30 16:29 AcroGfa4vE
-rw-r--r--   1 JDE      JDE          443 Jan 30 16:29 AcroFfa4vE
-rw-r--r--   1 JDE      JDE          443 Jan 30 16:29 AcroCfa4vE
-rw-r--r--   1 JDE      JDE      2147483647 Jan 31 02:53 AcroZoai0p
-rw-r--r--   1 JDE      JDE      2147483647 Jan 31 02:53 AcroP0aifq
-rw-r--r--   1 JDE      JDE      2147483647 Jan 31 02:53 AcroOOaOxq
-rw-r--r--   1 JDE      JDE      2147483647 Jan 31 02:53 AcrooxaOVr
-rw-r--r--   1 JDE      JDE      2147483647 Jan 31 02:53 AcroBOaOBr
-rw-r--r--   1 JDE      JDE      2147483647 Jan 31 03:05 AcroDfa4vE
-rw-r--r--   1 JDE      JDE          124 Jan 31 04:55 sendpdf.log
-rw-r--r--   1 JDE      JDE            0 Jan 31 04:56 Acro_JaW7Q
-rw-r--r--   1 JDE      JDE          443 Jan 31 04:56 Acro7JaW7Q
-rw-r--r--   1 JDE      JDE          443 Jan 31 04:56 Acro.JaW7Q
-rw-------   1 oracle   oinstall       0 Jan 31 04:58 croutHVPcYhGBq
-rw-rw-r--   1 root     sys          948 Jan 31 04:58 ps_data
-rw-r--r--   1 JDE      JDE      1421264 Jan 31 04:58 Acro8JaW7Q
Question by:matrix0511
Welcome to Experts Exchange

Add your voice to the tech community where 5M+ people just like you are talking about what matters.

  • Help others & share knowledge
  • Earn cash & points
  • Learn & ask questions
LVL 12

Expert Comment

by:Sharon Seth
ID: 39823640
I believe whatever is in /tmp should get deleted automatically once the process that created them exists
LVL 22

Expert Comment

ID: 39823808
It is not unusual for some processes to create temp files during normal processing. An application possibly called Acro something ? And running under the user "JDE".

Resist temptation to delete these files as long as the app is running.

/tmp is often a tmpfs filesystem (which means it can be stored in RAM) and shares space with the swap. There are two options to deal with your issue:

1. enlarge swap space
2. put /tmp on a regular filesystem

(2) will also fix the issue, but might cause performance degradation in the app because temp files are no longer stored in RAM/swap but on a regular disk.
LVL 21

Accepted Solution

tfewster earned 2000 total points
ID: 39824188
1) Check if the files are still in use with `fuser <filename>`; If that returns a process ID, they're still being held open so moving or deleting them won't release the space until the process ends. If they're not open, move them for further investigation or delete them.

2) Check if it's a recognisable file type with `file <filename>` and/or compare the first few bytes with a known file type. My guess would be that they're PDF files created by Acrobat writer output from JDE (JD Edwards?) processes, but they could be core dumps. Check that "sendpdf.log" file too - it's probably related and may have useful info. The "Acronnn" file created at 04:58 looks like a more normal sized output, though it's questionable why it wasn't tidied up when the creating process finished. Maybe another process cleans up the PDF files once they've been "sent" to their destination - e.g. emailed or archived.

3) Investigate cron and other schedulers to find what jobs should have been running at 02:53 (maybe starting at 02:00?) to prevent this happening again.

Given that all the large files are almost exactly 2GB, I'd guess the process hit a ulimit on the JDE user and failed. Question why the process should be writing such large files - e.g. a badly written report extract.

As you've tagged this questing with "Solaris", the issue of /tmp being shared with swap is quite likely, but I'd resist throwing more resources at the problem until you understand what's happening.
LVL 62

Expert Comment

ID: 39826245
Linux way (fuser on solaris unless you compile lsof)
lsof /tmp/*
lsof +L1

Author Comment

ID: 39826261
Thanks for your help!

Featured Post

Will your db performance match your db growth?

In Percona’s white paper “Performance at Scale: Keeping Your Database on Its Toes,” we take a high-level approach to what you need to think about when planning for database scalability.

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

Every server (virtual or physical) needs a console: and the console can be provided through hardware directly connected, software for remote connections, local connections, through a KVM, etc. This document explains the different types of consol…
Join Greg Farro and Ethan Banks from Packet Pushers (http://packetpushers.net/podcast/podcasts/pq-show-93-smart-network-monitoring-paessler-sponsored/) and Greg Ross from Paessler (https://www.paessler.com/prtg) for a discussion about smart network …
Connecting to an Amazon Linux EC2 Instance from Windows Using PuTTY.
Get a first impression of how PRTG looks and learn how it works.   This video is a short introduction to PRTG, as an initial overview or as a quick start for new PRTG users.
Suggested Courses
Course of the Month9 days, 15 hours left to enroll

762 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question