• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Napp-it - SCSI transport failed

T

TCMY95

n00b
Joined
Oct 9, 2014
Messages
20
Firstly I would just like to Say Gea napp-it is awesome and we love it. We have 2 storage appliances both running Napp-it but one of them is playing up intermittently. Periodically we get a SCSI transport failed: reason 'timeout': Giving up.

Here is our config in the server with issues;

Supermicro X9SRL-F
32GB Supermicro memory
LSI 9211-8i in IT mode
5 x WD red 4TB
1 x Samsung 128GB PRO - Read cache
1 x Intel 3500 100GB SSD - Write Cache

We boot from a pair of USB 3.0 32GB USB sticks 80MB/s read 80MB/s write.

I suspect it is an issue with the USB as the Iscsi interface stays up with no issues, it is only when we try to access the webui, logon through IPMI console or try to access an SMB share that we notice the issue.

Has anyone any thoughts on this?

thanks
 
For obvious reasons, check system log and fault log (menu system)
otherwise rule out usual problematic components like

- remove half of ram, check, try the other half
- remove usb sticks and install OS to any sata disk to boot
- use sata only (ahci)
 
Thanks Gea for your response.

My suspicions are the USB as we completed a memtest before the system went in, When we complete a warm reboot of the system it doesn't see a boot device. If we complete a cold reboot then the USB is detected again and the system boots normally.

Another thing we have noticed is the second USB boot device shows up as removed in the Napp-it console. What concerns me is that a. the USB devices are not mirrored and when we try to boot from one of the other USB slots the system doesn't boot.

I suppose the question is are we better booting from SATA drives. I thought booting from USB drives would be OK as the foot print of the system is so small but if this is going to cause reliability issues then we will replace the USB's.

thanks again..

Ian
 
Anyone

I have just looked at the logs and came across this kernal panic on multiple occasions.

ireport.os.sunos.panic.savecore_failure

thanks

ian
 
I have just adjusted the link aggregation on this box as I noted it was across 2 different cards and I was wondering if that was causing the issue. It only seems to happen under heavy network load.

Link Aggr was across

e1000g1
igb0
igb1

I have now reduced this to cover just the igb0 and 1. with the management address on e1000g0 and e1000g1 disabled.

fingers crossed.
 
_gea

On another note, once we have this stable we are looking to do some replication between 2 boxes and we had a quote from you not long ago. If we want to replicate from one box to another do we need one replication license or two? Also I noted that you provide support if we take the full package, can you offer direct support without the full package? for a price of course?

thanks

ian
 
You need only one key for the backup machine. It covers unlimited source servers.

Regarding support
I answer questions in forums but without guarantee of a reply.
This is because of limited time and the reason of limited options.
 
OK Thanks Gea. We will make a purchase as soon as I have confirmed this is stable, after every hard reboot it usually takes days before anything happens again. Do you think the link aggregation could cause the issue I have been experiencing?
 
Link aggregation can give performance and can cause problems.
On problems it is one of the first things that I would switch off.

I do not use aggregation and prefer 10G all the way
 
Hi gea

We are going to take your advice. on the 10GB. Do you know if Mellanox cards work as they do not list Solaris. However the card we have in the other storage device isn't list either (it is Intel 10GB) but that one works?
 
If your using iscsi, I would not use link aggression, but just let it do multipathing instead on your initiators.
 
thanks Patrickdk. I am fairly new to the ZFS territory but I am having to learn very fast as we have had some issues with one of the appliances. Anyway am I right in thinking that if we assign different IP addresses to the individual network ports citrix xen server will recognise them as part of the same LUN and apply multipathing?
 
I have no idea how citrix works.

But it normally noticed the source machine and serial of the disks are the same, and it will multipath it. On linux this is a semi-manual process to setup multipath, on vmware it just works, and on windows it normally just works.

You will have to pick if you want to use multible ip's, one per nic, on the same subnet, or on different subnets. Normally using seperate subnets is perfered, but also makes it more annoying to setup.
 
IGC said:
Hi gea

We are going to take your advice. on the 10GB. Do you know if Mellanox cards work as they do not list Solaris.
Click to expand...

Try it.
It will probably work out of the box - or never:
But prefer Intel for 10Gb Ethernet (I use nothing else) !!!
 
decided on intel, it works fine in the other machine and in the grand scheme of things a few extra ££ is worth the hassle free experience.
 
You must log in or register to reply here.
Back
Top