Hi
I have an IBM blade with internal harddisks, in hardware RAID1. I've installed a CentOS 6 64bit on it, everything works just fine.
After the installation, I've presented a vdisk to the blade from an external SAN (an HP EVA4000), connected through FiberChannel. I've partitioned the disk, formatted it and mounted it under /store, then added it to fstab. Everything was fine, until I rebooted.
At boot I'm getting the following error for /store:
Mounting local filesystems: mount: special device UUID=2a587e95-4a6c-4336-bb8b-f0d066905bc5 does not exist
It just goes on to boot without mounting this filesystem. After it boots, I can log in and give the command "mount -a", and it gets mounted without problems.
As far as I can tell, the reason for this is that CentOS doesn't wait for the external disk to get initialized fully and it just doesn't find it at boot time. I have other CentOS blades, installed and booting from the same SAN and they work without problems, but I noticed that they wait a little bit longer at boot.
Am I missing some stuff from initrd? What can I do to make it wait for the block device a bit longer before it tries mounting it ?
Thanks!
Does it work usin netdev option?
Eero 10.10.2015 4.17 ip. "Imre Gergely" gimre@narancs.net kirjoitti:
Hi
I have an IBM blade with internal harddisks, in hardware RAID1. I've installed a CentOS 6 64bit on it, everything works just fine.
After the installation, I've presented a vdisk to the blade from an external SAN (an HP EVA4000), connected through FiberChannel. I've partitioned the disk, formatted it and mounted it under /store, then added it to fstab. Everything was fine, until I rebooted.
At boot I'm getting the following error for /store:
Mounting local filesystems: mount: special device UUID=2a587e95-4a6c-4336-bb8b-f0d066905bc5 does not exist
It just goes on to boot without mounting this filesystem. After it boots, I can log in and give the command "mount -a", and it gets mounted without problems.
As far as I can tell, the reason for this is that CentOS doesn't wait for the external disk to get initialized fully and it just doesn't find it at boot time. I have other CentOS blades, installed and booting from the same SAN and they work without problems, but I noticed that they wait a little bit longer at boot.
Am I missing some stuff from initrd? What can I do to make it wait for the block device a bit longer before it tries mounting it ?
Thanks!
-- Imre Gergely http://havaz.net gpg --keyserver subkeys.pgp.net --recv-keys 0x34525305
CentOS mailing list CentOS@centos.org https://lists.centos.org/mailman/listinfo/centos
_netdev The filesystem resides on a device that requires network access (used to prevent the system from attempting to mount these filesystems until the network has been enabled on the system).
This device is not a network device (this a SAN not a NAS). To the OS it looks like a normal SCSI attached device, it's /dev/sdb. In the blade there is a HBA (Qlogic) card, and it's connected through FiberChannel. If I understand these terms correctly, it has nothing to do with the network.
On 10/10/2015 04:20 PM, Eero Volotinen wrote:
Does it work usin netdev option?
Eero 10.10.2015 4.17 ip. "Imre Gergely" gimre@narancs.net kirjoitti:
Hi
I have an IBM blade with internal harddisks, in hardware RAID1. I've installed a CentOS 6 64bit on it, everything works just fine.
After the installation, I've presented a vdisk to the blade from an external SAN (an HP EVA4000), connected through FiberChannel. I've partitioned the disk, formatted it and mounted it under /store, then added it to fstab. Everything was fine, until I rebooted.
At boot I'm getting the following error for /store:
Mounting local filesystems: mount: special device UUID=2a587e95-4a6c-4336-bb8b-f0d066905bc5 does not exist
It just goes on to boot without mounting this filesystem. After it boots, I can log in and give the command "mount -a", and it gets mounted without problems.
As far as I can tell, the reason for this is that CentOS doesn't wait for the external disk to get initialized fully and it just doesn't find it at boot time. I have other CentOS blades, installed and booting from the same SAN and they work without problems, but I noticed that they wait a little bit longer at boot.
Am I missing some stuff from initrd? What can I do to make it wait for the block device a bit longer before it tries mounting it ?
Thanks!
-- Imre Gergely http://havaz.net gpg --keyserver subkeys.pgp.net --recv-keys 0x34525305
CentOS mailing list CentOS@centos.org https://lists.centos.org/mailman/listinfo/centos
CentOS mailing list CentOS@centos.org https://lists.centos.org/mailman/listinfo/centos
On October 10, 2015 8:34:11 AM CDT, Imre Gergely gimre@narancs.net wrote:
_netdev The filesystem resides on a device that requires network access (used to prevent the system from attempting to mount these filesystems until the network has been enabled on the system).
This device is not a network device (this a SAN not a NAS). To the OS it looks like a normal SCSI attached device, it's /dev/sdb. In the blade there is a HBA (Qlogic) card, and it's connected through FiberChannel. If I understand these terms correctly, it has nothing to do with the network.
On 10/10/2015 04:20 PM, Eero Volotinen wrote:
Does it work usin netdev option?
Eero 10.10.2015 4.17 ip. "Imre Gergely" gimre@narancs.net kirjoitti:
Hi
I have an IBM blade with internal harddisks, in hardware RAID1. I've installed a CentOS 6 64bit on it, everything works just fine.
After the installation, I've presented a vdisk to the blade from an external SAN (an HP EVA4000), connected through FiberChannel. I've partitioned the disk, formatted it and mounted it under /store, then added it to fstab. Everything was fine, until I rebooted.
At boot I'm getting the following error for /store:
Mounting local filesystems: mount: special device UUID=2a587e95-4a6c-4336-bb8b-f0d066905bc5 does not exist
It just goes on to boot without mounting this filesystem. After it boots, I can log in and give the command "mount -a", and it gets
mounted
without problems.
As far as I can tell, the reason for this is that CentOS doesn't
wait
for the external disk to get initialized fully and it just doesn't
find
it at boot time. I have other CentOS blades, installed and booting
from
the same SAN and they work without problems, but I noticed that they wait a little bit longer at boot.
Am I missing some stuff from initrd? What can I do to make it wait
for
the block device a bit longer before it tries mounting it ?
Does dmesg provide any useful information? If you remove (assuming they are there) rhgb and quiet and add debug to the kernel line at boot time does that give you any more info? How does the dmesg output compare to one that is able to mount the disk at boot without issue?
Barry
On 10/10/2015 05:03 PM, Barry Brimer wrote:
On October 10, 2015 8:34:11 AM CDT, Imre Gergely gimre@narancs.net wrote:
_netdev The filesystem resides on a device that requires network access (used to prevent the system from attempting to mount these filesystems until the network has been enabled on the system).
This device is not a network device (this a SAN not a NAS). To the OS it looks like a normal SCSI attached device, it's /dev/sdb. In the blade there is a HBA (Qlogic) card, and it's connected through FiberChannel. If I understand these terms correctly, it has nothing to do with the network.
On 10/10/2015 04:20 PM, Eero Volotinen wrote:
Does it work usin netdev option?
Eero 10.10.2015 4.17 ip. "Imre Gergely" gimre@narancs.net kirjoitti:
Hi
I have an IBM blade with internal harddisks, in hardware RAID1. I've installed a CentOS 6 64bit on it, everything works just fine.
After the installation, I've presented a vdisk to the blade from an external SAN (an HP EVA4000), connected through FiberChannel. I've partitioned the disk, formatted it and mounted it under /store, then added it to fstab. Everything was fine, until I rebooted.
At boot I'm getting the following error for /store:
Mounting local filesystems: mount: special device UUID=2a587e95-4a6c-4336-bb8b-f0d066905bc5 does not exist
It just goes on to boot without mounting this filesystem. After it boots, I can log in and give the command "mount -a", and it gets
mounted
without problems.
As far as I can tell, the reason for this is that CentOS doesn't
wait
for the external disk to get initialized fully and it just doesn't
find
it at boot time. I have other CentOS blades, installed and booting
from
the same SAN and they work without problems, but I noticed that they wait a little bit longer at boot.
Am I missing some stuff from initrd? What can I do to make it wait
for
the block device a bit longer before it tries mounting it ?
Does dmesg provide any useful information? If you remove (assuming they are there) rhgb and quiet and add debug to the kernel line at boot time does that give you any more info? How does the dmesg output compare to one that is able to mount the disk at boot without issue?
Yes, I tried that, and look at dmesg, but couldn't find anything interesting. The only useful info is that the drive is not ready when it tries to mount it, but it becomes ready a little bit later. The one that is mounting the disk without issue is a bit different because it also boots from the SAN so it really needs to have the external storage available from the beginning. This one boots from internal disk and only later wants to mount the external disk.
On 10/10/2015 11:41 AM, Imre Gergely wrote:
The one that is mounting the disk without issue is a bit different because it also boots from the SAN
...which means that the HBA driver is included in the initrd, but not in the system where you're having trouble.
Edit /etc/dracut.conf.d/hba.conf and add one line: add_drivers+="your_hba_driver"
On Oct 10, 2015, at 9:34 AM, Imre Gergely gimre@narancs.net wrote: _netdev The filesystem resides on a device that requires network access (used to prevent the system from attempting to mount these filesystems until the network has been enabled on the system).
_netdev in fstab was a workaround from oracle linux support for a FC issue very similar to the one you described(OS/driver may not plogi in to the storage quickly enough.). At least give it try before so quickly dismissing assistance.
On 10/10/2015 10:06 PM, Steven Tardy wrote:
On Oct 10, 2015, at 9:34 AM, Imre Gergely gimre@narancs.net wrote: _netdev The filesystem resides on a device that requires network access (used to prevent the system from attempting to mount these filesystems until the network has been enabled on the system).
_netdev in fstab was a workaround from oracle linux support for a FC issue very similar to the one you described(OS/driver may not plogi in to the storage quickly enough.). At least give it try before so quickly dismissing assistance. _______________________________________________
You are right of course. I'm sorry. I will try it tonight and let you know.
-
Barry Brimer -
Eero Volotinen -
Gordon Messmer -
Imre Gergely -
Steven Tardy