Brian, Ryan, I've heard of this happening once or twice recently. I can't find the exact thread/github issue where I read about it but I feel like I remember seeing that this was a bug with VirtualBox itself (which sounds right to me) and that they confirmed/fixed it but that the fix hasn't been released yet.
You MIGHT see a fix if you downgrade VirtualBox a patch release or two. But honestly, I'm not sure. best, Mitchell On Mon, Feb 3, 2014 at 12:01 PM, Brian Lalor <[email protected]> wrote: > I've also had this problem, using the same base box. Ryan and I have been > using the base box for well over a year and the problems are relatively > new. Are you sure the guest additions are 4.3.6? Are you updating them > yourself? The ones baked into the box are fairly old. > > I'm wondering if it could be related to newer versions of VirtualBox. If > the load's climbing but top doesn't show anything, it could be the kernel > spinning out of control. Top will show you the amount of time the CPU's > spending in user/system/io wait/idle time; it'd be interesting to see where > it's spending its time. iotop is useful for finding the culprit, even if > it's a kernel thread. > > On Feb 2, 2014, at 9:42 PM, Ryan Doherty <[email protected]> wrote: > > Hey everyone, been googling and troubleshooting this issue for days, > hoping to get some ideas on where else to look to fix this problem. > > After I boot up my vagrant image (CentOS release 6.3 (Final)), about 30 > minutes later all SSH connections stop working and my Rails dev server > launched inside of the vagrant image grinds to a halt (stops responding too > via http). Any attempts to create a new ssh connection to it results in > this error: > > ssh_exchange_identification: Connection closed by remote host > > If I boot up in GUI mode the same problem occurs for ssh connections made > via `vagrant ssh`, but the VirtualBox GUI terminal still responds. However, > it prints out this error message every few minutes: > > INFO: task events/1:12 blocked for more than 120 seconds. > "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. > INFO: task sshd:2878 blocked for more than 120 seconds > "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. > > I ran top to see if anything was suspicious, nothing was taking up CPU, > but the load averages continually increased from ~0.5 to >2 over a period > of 5 minutes and I'm guessing would have kept increasing if I didn't kill > the machine. > > I have to `vagrant halt -f` to stop the machine, otherwise it will > eventually close all ssh connections after about 10 more minutes. When I > `vagrant up` after everything works fine again for about 30 minutes. > > Vagrant 1.4.3 > Guest Additions 4.3.6 > OSX 10.9.1 > > I've found a lot of ssh problems and solutions elsewhere, but this one > seems different in that ssh works fine for ~30 minutes then stops working. > Thanks in advance for any advice on what else to try! > > -Ryan > > -- > You received this message because you are subscribed to the Google Groups > "Vagrant" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/groups/opt_out. > > > -- > Brian Lalor > [email protected] > > -- > You received this message because you are subscribed to the Google Groups > "Vagrant" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/groups/opt_out. > -- You received this message because you are subscribed to the Google Groups "Vagrant" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
