• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Virtualization Issue on Ubuntu 14.04

genoshan

Junior Member
Hi forum! I didn't find an Introduction Thread , so I'll introduce myself now.
I'm a big enthusiast of technologys most of it the hardware aspect of IT.
Anyway I have a rol as a sysadmin at the job and have been running in to difficulties as of late.
I have a question about virtualization on an Hyper-v environment and Ubuntu 14.04.
The matter is that from time to time, the systems get "corrupted" and puts the entyre FileSystem on read-only mode, disabling everything the system is up to.
I have checked that the linux integration services are installed, verifying with the command lsmod . :colbert:
Hase anyone a clue on this issue?
 
Depending on what you are using for the storage solution on your Hyper-V cluster, I would take a look at the error logs (both at the Hyper-V and SAN level). I'm not familiar with Hyper-V, but in other virtualization solutions (read: VMware), you can view log files and see where any disconnects or potential corruption is coming from.

The fact your guest VM(s) are being thrown into a read-only state, indicates there's an issue with Hyper-V talking to your SAN. I would double check your network links to each virtualization host, your Hyper-V "head" and the SAN.
 
Depending on what you are using for the storage solution on your Hyper-V cluster, I would take a look at the error logs (both at the Hyper-V and SAN level). I'm not familiar with Hyper-V, but in other virtualization solutions (read: VMware), you can view log files and see where any disconnects or potential corruption is coming from.

The fact your guest VM(s) are being thrown into a read-only state, indicates there's an issue with Hyper-V talking to your SAN. I would double check your network links to each virtualization host, your Hyper-V "head" and the SAN.

I've seen this kind of thing happen before on REAL hardware before too. For what it's worth, you can't rule out physical links going down, but from what I've seen, it's usually an underlying driver issue. Basically, what happens is the file system is doing what it does...it starts getting backlogged due to write errors or cache errors and things start to unravel.

1. Read the Logs - I'd start by looking at system logs to try to isolate the problem.
2. Physical Storage Network Layer - You could be dealing with a SAN connection issue if your HBAs aren't on their latest firmware or if you're dealing with iSCSI, check network statistics to see if you have any evidence of problems there.

Another thing to consider is limited disk I/O. It's not uncommon in virtualized environments to carve up a 1-2TB LUN and connect 30 VMs to it... All it takes is hitting the threshold where there's too much chatter on the LUN and you've spent your disk I/O. the only solution is to get more spindles or switch to a block-level SAN solution with storage tiering or SSD.

3. Check HyperV and see if there are OTHER virtual storage adapters to use...or other drivers Ubuntu can use for virtualization. As I stated before, drivers are possibly a cause here. Some drivers and virtual hardware work better with virtualization simply because they are more patient and don't timeout as quickly. It's not uncommon for resource contention to cause slowdowns and some drivers give a good cushion to the disk...

4. Finally, look at your Ubuntu swap configuration. This may be something you can tune to take some load off the disk. You can possibly throw more ram and less swap at the system or check best practices...I've not played with HyperV much. I don't know all the sweet spots, but there are always caveats to virtualization and some work better than others with certain OSes. I have had no problems with VMware and linux....or VirtualBox and linux. (Red Hat, Cent, SuSE, and Ubuntu flavors)
 
Hyper-v 3.0 (2012 and 2012R2) and Ubuntu 13+ work very well together. We have had no problems with our internal systems that run on Ubuntu.

As mentioned above start looking at the logs on both the VM, hypervisor and storage.
 
Thank you all for your reply, Since I'm not in specifically in charge of the storage I will try to get the corresponding logs to check anything that is interfering with the correct communication.
 
Have you checked your hypervisor's logs yet as folks mentioned? It also depends on your actual storage type. Do you have fiber or iSCSI?

For instance we had a fiber SAN. When we experienced fiber hiccups I could search my hypervisor logs for anything qla"xxx"(fiber channel stuff) related to determine any pathing issues(timeouts or dropouts) or SAN problems from the perspective of the hypervisor.

I have experienced your issue before on our platform.

The simplest explanation for what could be happening is that your hypervisor and SAN are periodically having connection drops or timeouts. Multipathing can help but you probably have other SAN issues going on too. It is hard to help further without some more info from you.

While I am not familiar with Hyper-V as well I would not think you setup multipathing from the guest(Ubuntu) as your traffic goes through your hypervisor host. You would set it up from the hypervisor host.
 
Thank you RearAdmiral for your response. I do not have that info with me right now, since I install the ubuntu systems only, but don't have the network info.
Which are the questions I need to ask to the Network team?, in order to tell them where to look exactly, and not to take too much time from them, since they are always busy.
Regards!
 
Thank you RearAdmiral for your response. I do not have that info with me right now, since I install the ubuntu systems only, but don't have the network info.
Which are the questions I need to ask to the Network team?, in order to tell them where to look exactly, and not to take too much time from them, since they are always busy.
Regards!

Given how you mention network it sounds like you do not have a fiber channel SAN/Fabric. I would recommend not just throwing out a random request or question to the team. They would like to see any logs you can provide from the hypervisor or VM level from the time when the VM went into RO mode. You gotta give them something. You still need to do your homework / log collection to help these guys resolve any possible network hiccups.
 
Back
Top