3.8.13 can not boot

hi all,

   I install the 3.8.13 rockstor ,first I come in the system,the rockstor service is not up,I use command "systemctl start rockstor " it can up ,but when I reboot the system can not boot ??????
   I find the progress all in "starting preprocess NFS configuration" ,can not go on..........

@ygd2013520 Is this a fresh install from the Rockstor iso, and does it happen on every boot? I have seen very infrequently a looping behaviour that looks like some kind of systemd race condition between 3 services during boot (non of them were rockstor services) but was not able to reproduce it.

Thanks for reply. YES,it happened on every boot, then I select the “rescue mode” to boot, it can boot sometimes,but most of times can not boot.I use this new ISO on 3 PCs, they are all the same(can not boot).Maybe it is a bug in 3.8.13 ISO ?

Hi @phillxnet ,some times it can boot, but when I reboot it go to this step and can not continue.Please see it in the picture.

Maybe this is related to an issue I opened:

My Systems hang on nearly the same point. And I can not get a single system to boot reliably. Ich have tried different systems with different hardware.
The only system on which I did not run into this problem is a VM. On real hardware I can not get rockstor to boot reliably.

Thanks for reply. I If you got the method ,please tell me,thank you very much.

I get this too on physical hardware. Fails to boot perhaps 80% of the time on my two NASes, a HP MicroServer Gen8 G1610T and a self-build. Strangely it seems more reliable when selecting the debug boot option, then press e on keyboard to edit the boot options and remove the “quiet” parameter from the second-last line (then Ctrl-X to boot). Also it seems to do better when IPMI or HP iLO out-of-band remote control is not active, so I disconnect these during boot. Seems to be more like 50% success with this approach which I’ve been using ever since I started using Rockstor around six months ago. When it fails, it gets to a different message in the debug boot log each time. Would be great to get this fixed.

Hi all, maybe a stupid question, but i ask - because not all explicit told this: all of you running with nfs and nfs mount on fstab?

Why do i ask this? Once Got same issue - not on Rockstor, on a debian - after a kernel upgrade

Thanks for reply to ,I will try your method.

I am also having this same problem on a fresh install. Just a blinking cursor to look at on boot.

Not NFS-related as I’ve disabled that and still can’t boot. Doesn’t happen on a freshly installed Rockstor with no pools\shares created. I think it’s a race\timing issue with particular btrfs subvols.

Hi all,

I finally found a solution for this on my system. I seems that it works fine now.

It seems the problem is NFS-related, but it is not directly related to nfs.service. So disabling this does not help.
The system hangs on nfs-config.service. And this is before nfs.service.

I modified /usr/lib/systemd/system/nfs-config.service
I modified the line:
After=local-fs.target

to this:
After=local-fs.target network.target

Please try if this solution works for you too and report it here.
I think, nfs-config.service is simply startet to early. I came to this, because booting works reliably when network cable is disconnected.

2 Likes

Didn’t make any difference in my experience, still fails to boot most of the time. I take it disabling the NFS service doesn’t prevent the nfs-config.service starting then? If not, is there a way to prevent the nfs-config.service starting at all? That would at least establish whether this is the culprit.

My Rockstors have 20TB & 16TB of storage, appx 60% utilised. I’m wondering whether the devs have tested the boot process at this scale.

Does your system boot reliably when network cable is disconnected?

My system boots reliably with the above modification to the nfs-config.service.