• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

Intermittent error: filemng failed: Agent.php Line 243

danallen1

Basic Pleskian
Plesk version: 12.0.18 Update #64 , last updated at Sept 9, 2015 07:16 PM
Ubuntu 14.04
This server was created on Sept 9, 2015 and has very little activity. Nothing installed except Ubuntu and Plesk. One domain.


Just now, I ran 20 clicks of the FILES tab and got errors on 8 of them. Error looks like this:

upload_2015-9-15_23-4-29.png

Plesk version:
12.0.18 Update #64 , last updated at Sept 9, 2015 07:16 PM



Internal error: filemng failed:
Message filemng failed:
File Agent.php
Line 243
Type PleskUtilException
Go To Previous Page
This is a brand new server, created to fix this problem that was occurring before the rebuild. When I say rebuild, here are the steps I took:
  1. Spun up a new vps at Digital Ocean with Ubuntu 14.04
  2. Set the fully qualified host name, as required for the Plesk installation to work.
  3. Downloaded the plesk installer (wget http://autoinstall.plesk.com/plesk-installer)
  4. Made the installer executable (chmod +x plesk-installer)
  5. Ran the installer (./plesk-installer ) (see below for config)
  6. Loaded the files for the one domain that runs on that server
  7. Everything was good, until........
  8. I received a report from the system administator that she is unable to access the FILES tab, as a result of getting the error noted above
  9. I logged in and got the error 8 out of 20 times clicking the FILES tab
The VPS provider, DigitalOcean, refers to VPSs as "droplets". Here is a summary of its statuys and historyL:
upload_2015-9-15_23-16-6.png

upload_2015-9-15_23-10-17.png

upload_2015-9-15_23-10-34.png

upload_2015-9-15_23-11-8.png
 
A ton was sent to the log just now, on a single click of the file tab. Full monty is attached, just the start below.

Ah ha! Not enough memory? Highlighted below in the first line from the log

I could be mistaken but I thought there was a swap file on the previous server.

Is there anything we can do besides add memory?




[16-Sep-2015 17:58:15 America/New_York] PHP Warning: proc_open(): fork failed - Cannot allocate memory; File: /opt/psa/admin/plib/Service/Agent/Transport/Util.php, Line: 27

[16-Sep-2015 17:58:15 America/New_York] Exception: PHP Warning: proc_open(): fork failed - Cannot allocate memory; File: /opt/psa/admin/plib/Service/Agent/Transport/Util.php, Line: 27​
 

Attachments

  • panellog.txt
    42 KB · Views: 1
yes, it is a vps.

Does this situation indicate the requirements document might need to be updated to show a higher minimum memory requirement?
 
Hi danallen1,

Does this situation indicate the requirements document might need to be updated to show a higher minimum memory requirement?

well... yes and no.
The minimum requirements as listed at "http://www.odin.com/products/plesk/requirements/" don't include extensions usage and are based on a real "minimum" environment, as for example: minimum Plesk installation, 1 domain only and no additional services in the background, which as well consume RAM. Here is an external link for a pretty good tutorial, which might help you to investigate your RAM usage:

http://www.cyberciti.biz/faq/linux-check-memory-usage/ ( external link, please inform me, if the link goes dead, so I can replace the link with another one )
 
higher minimum memory requirement?
VPS has many resources that may be considered as "different types of memory". And if there is lack of one of these resources - "Cannot allocate memory" may occur.
For instance on Virtuozzo VPS:

# cat /proc/user_beancounters
Version: 2.5
uid resource held maxheld barrier limit failcnt
102: kmemsize 15499821 20713472 9223372036854775807 9223372036854775807 0
lockedpages 0 4 131072 131072 0
privvmpages 242856 266779 9223372036854775807 9223372036854775807 0
shmpages 34822 34838 9223372036854775807 9223372036854775807 0
dummy 0 0 9223372036854775807 9223372036854775807 0
numproc 96 114 9223372036854775807 9223372036854775807 0
physpages 104740 131088 131072 131072 0
vmguarpages 0 0 262144 262144 0
oomguarpages 75818 76198 131072 131072 0
numtcpsock 34 42 9223372036854775807 9223372036854775807 0
numflock 14 40 9223372036854775807 9223372036854775807 0
numpty 1 1 9223372036854775807 9223372036854775807 0
numsiginfo 0 21 9223372036854775807 9223372036854775807 0
tcpsndbuf 592960 993432 9223372036854775807 9223372036854775807 0
tcprcvbuf 557056 1007328 9223372036854775807 9223372036854775807 0
othersockbuf 182648 244480 9223372036854775807 9223372036854775807 0
dgramrcvbuf 0 30912 9223372036854775807 9223372036854775807 0
numothersock 137 156 9223372036854775807 9223372036854775807 0
dcachesize 5664825 10086364 9223372036854775807 9223372036854775807 0
numfile 1549 1715 9223372036854775807 9223372036854775807 0
dummy 0 0 9223372036854775807 9223372036854775807 0
dummy 0 0 9223372036854775807 9223372036854775807 0
dummy 0 0 9223372036854775807 9223372036854775807 0
numiptent 20 20 9223372036854775807 9223372036854775807 0

All *pages are memory related resources. And if failcnt is not 0 - there is lack of this resource.
 
well... yes and no.
I think Odin might want to do something about this.

I figured Odin was the end of Plesk being cool, but Odin has been doing a great job from what I can see. Given that Odin is doing a good job of doing the right things, I think Odin might want to update guidance, to say that more than 512M RAM on a VPS is required if unless you can live with occasional bombs of the Plesk web app, that do not cause damage, but do terminate some processes. This notice could be made prominent in the installation process. That is just my take, i know I am a moron.

VPS has many resources that may be considered as "different types of memory". And if there is lack of one of these resources - "Cannot allocate memory" may occur.
For instance on Virtuozzo VPS:

# cat /proc/user_beancounters
Version: 2.5
uid resource held maxheld barrier limit failcnt
102: kmemsize 15499821 20713472 9223372036854775807 9223372036854775807 0
.
.
.if failcnt is not 0 - there is lack of this resource.
I appreciate this background a lot, thank you.

Question:
Would there be a chance the file tab bomb I reported to start this thread would disappear if the RAM on the server increased from 512M to 1G?
 
Back
Top