I've just migrated my PRTG installation from one VPS (Virtual Private Server) hosting company to another server and across all the devices and sensors I now get alarms every 1 minute or so from one or more of the 77 sensors with "Access is Denied (0x5)". The old server is still live and does not get same error.
Access is Denied (0x5)
Votes:
0
Best Answer
Accepted Answer
Votes:
0
Short answer: You are quite likely overloading your VPS system!
Long answer:
- Are you aware that using a cheap VPS system is potentially not a good choice for running a monitoring system? Most low-budget Windows VPS run on Virtuozzo ("Parallels Virtuozzo Containers") and most hosters run far too many "Containers" on their system to save money. These containers can affect each other much more than in a "real" virtualization (XEN, VMware) system. If you see longer request times in your monitoring or if you see errors these could be caused by other users or simply overloading of the host server.
- The "Access Denied" could be triggered by a resource-limiting setting of your container (most often the cause is: Not enough memory).
- Are you aware that the HTTP Full Page sensor actually starts an full scale copy of the locally installed Internet Explorer, opens the webpage, measures the time to load the whole page (with all assets, even Flash files and other resource intensive stuff), and then closes IE? The whole process can be quite resource intensive as you can imagine. IE is quite "resource wasting". You have 74 sensors of these with 1 second interval => more than 1 Internet Explorer per second is started, loads a page and quits. This could even kill a real hardware system with weak performance. I would not be surprised if a cheap VPS system is overloaded with just a third of this load.
- Do you really need full page sensors for each URL every 60 seconds? Consider using a standard HTTP sensor even 60 seconds (it requires only an estimated 1-5% of the CPU cycles) and will still give you a warning if the server fails. Then run a full page sensor for each URL every 15 minutes if you need the full page feature, too.
- What version of IE is installed on your VPS? We recommend running IE8.
Please consider moving your monitoring system to a serious cloud hoster. We recommend Amazon EC2 and RackSpaceCloud (their smalles Windows machine is below US$60 per month: http://www.rackspacecloud.com/cloud_hosting_products/servers/pricing/). We use both a lot ourselves and PRTG runs great on both platforms.
Created on Feb 5, 2011 8:17:19 PM by
5 Replies
Votes:
0
Hello,
on which exact sensors do you get this error? Are the 77sensors all sensors? Or are there other sensors from the same type showing no errors? What are the problematic sensors monitoring?
Best Regards.
Created on Feb 1, 2011 2:09:52 PM by
Votes:
0
Could it be that the User Account under which the PRTG Probe runs makes a difference? If this is set to "LOCAL SYSTEM" can you try if changing it to a Admin user brings improvement?
New scanning intervals can be configured under "Setup"->"System Administration"->"System & Website" in PRTG.
Created on Feb 1, 2011 3:46:04 PM by
Accepted Answer
Votes:
0
Short answer: You are quite likely overloading your VPS system!
Long answer:
- Are you aware that using a cheap VPS system is potentially not a good choice for running a monitoring system? Most low-budget Windows VPS run on Virtuozzo ("Parallels Virtuozzo Containers") and most hosters run far too many "Containers" on their system to save money. These containers can affect each other much more than in a "real" virtualization (XEN, VMware) system. If you see longer request times in your monitoring or if you see errors these could be caused by other users or simply overloading of the host server.
- The "Access Denied" could be triggered by a resource-limiting setting of your container (most often the cause is: Not enough memory).
- Are you aware that the HTTP Full Page sensor actually starts an full scale copy of the locally installed Internet Explorer, opens the webpage, measures the time to load the whole page (with all assets, even Flash files and other resource intensive stuff), and then closes IE? The whole process can be quite resource intensive as you can imagine. IE is quite "resource wasting". You have 74 sensors of these with 1 second interval => more than 1 Internet Explorer per second is started, loads a page and quits. This could even kill a real hardware system with weak performance. I would not be surprised if a cheap VPS system is overloaded with just a third of this load.
- Do you really need full page sensors for each URL every 60 seconds? Consider using a standard HTTP sensor even 60 seconds (it requires only an estimated 1-5% of the CPU cycles) and will still give you a warning if the server fails. Then run a full page sensor for each URL every 15 minutes if you need the full page feature, too.
- What version of IE is installed on your VPS? We recommend running IE8.
Please consider moving your monitoring system to a serious cloud hoster. We recommend Amazon EC2 and RackSpaceCloud (their smalles Windows machine is below US$60 per month: http://www.rackspacecloud.com/cloud_hosting_products/servers/pricing/). We use both a lot ourselves and PRTG runs great on both platforms.
Created on Feb 5, 2011 8:17:19 PM by
Disclaimer: The information in the Paessler Knowledge Base comes without warranty of any kind. Use at your own risk. Before applying any instructions please exercise proper system administrator housekeeping. You must make sure that a proper backup of all your data is available.
Add comment