Unable to login to web ui

C

Caliban

New Member
Apr 4, 2011
5
0
1
We had a power outage last week, and the UPS failed (of course). All VMs appear to have come back online after power was restored, so I didn't think much of it. Today however, I tried logging into the webui, and at first it couldn't even connect. Turned out apache wasn't started (why? nothing in the logs about this). After manually starting it, the ui shows up, but it won't accept my password. A little investigating showed this in the daemon.log:
"login failure: 500 Can't connect to 127.0.0.1:83 (connect: Connection refused)"

What is running on port 83?

I did poke around a little, and restarted pvedaemon, pvenetcommit and pvetunnel, and now, when I try to login, the above error message is replaced by:
"PAM auth failed: User not known to the underlying authentication module"

I don't want to take any risks, and am planning to install the latest pve on a different machine, and migrate the VMs to that, but in order to do that, it would be helpful if I could at least login to this one...

Any ideas how I can persuade it to let me log in?

Thanks,
Peter
 
what kind of storage system do you get? raid controller with BBU? and did you disable the hard drive cache (I mean the cache in the hdd´s, not the raid controller cache).

if files are gone after a power loss its mainly due to missing cache protection of the storage system.

back to your question:
can you login via ssh/console? if yes, try changing the password with passwd and try again on the web.
 
Thanks for the quick response Tom!

There is no BBU on the raid controller unfortunately, or if there is, the battery is obviously dead (it a rather old beast, so that wouldn't surprise me at all).

I'm not sure how to disable the cache on the drives though, nor am I sure if I'm willing to go that route, since it seems to be a simple matter of a corrupt file, and/or a daemon not starting or something.

I am able to ssh into it, and since I'm using the same username/password for that, resetting it would not do anything, right? I've been digging through all sorts of logfiles to find a mentioning of why Apache did not start on bootup, or whatever daemon is supposed to listen on port 83, but can't find anything of the sort...

There's no huge emergency yet, so I have time to analyze what happened, but I'm kind stumped...
 
Caliban said:
Thanks for the quick response Tom!

There is no BBU on the raid controller unfortunately, or if there is, the battery is obviously dead (it a rather old beast, so that wouldn't surprise me at all).

I'm not sure how to disable the cache on the drives though, nor am I sure if I'm willing to go that route, since it seems to be a simple matter of a corrupt file, and/or a daemon not starting or something.
...
Click to expand...

my hints will prevent corrupt files...

try re-installing the pve packages.
 
Thanks Dietmar. That explains why things changed when I started that. Still find it odd why it didn't start on bootup though. And yes, using username root.

I also noticed that my nightly backups aren't running anymore. Probably another daemon that I have to start manually (which one?)... What else failed to start?

Anyway, things are starting to look a little brighter. Today I tried to login to the ui again, and low an behold, it let me in! Somehow it fixed itself overnight, or at least part of itself. So at least I'm on my way to migrating VMs to another node. It's all very strange though, and would love to get to the bottom of it. If nothing else, for my own education:)

@Tom: If I read your response correctly, I am supposed to disable the cache on the HDDs to prevent corrupt files? Interesting... I will read more on that. Thanks for the lead!
 
Interesting read... It makes sense, and it definitely something I didn't know!

Thanks again!
 
Just an update. I found that crond wasn't running, which is responisble for backups, so that part is fixed as well.

So, without having to dig too deep into proxmox's internal workings, I'm a little confused about the device names. I assume I need to target /dev/hda if I want to disable caching? If that's the case I should be good, since I get:
Code: 
# hdparm -W /dev/hda

/dev/hda:
 write-caching = not supported
Would that translate to both drives though (hardware RAID 1)? If I need to target another device(s), please enlighten me, because I get lost in the device mappings going on.

Edit: Maybe I should stop asking questions before I finish my coffee. I need to set this using software that came with the controller, right, not hdparm?
 
Last edited:
You must log in or register to reply here.
Top Bottom