OpenSolaris derived ZFS NAS/ SAN (OmniOS, OpenIndiana, Solaris and napp-it)

ZFS will not trust the data on the old disk but does a full checksum check with an optional repair from redundancy. A resilver and a replace can take a similar amount of time.

Your %done and estimated hours togo is what you need to check.

ps
A new disk must not be initialized when there is no ZFS label on it or it is partinioned. A NTFS disk with a single partition is perfect. A "no such device" means that the disk is not detected correctly, maybe a unstable cabling or power can be the reason.
 
Last edited:
Thanks, it took 32 hours in the end. I haven't had a disk failure yet, so now I know what to expect.

I'll do the next ones 2 by 2.
 
Another related question : the replace is going on, and I'm writing to the pool. Does the old drive get updated ?
 
It is obvious that if your "replace from" disk was online, that it got updated as a failure of the ""replace to" disk will otherwise reduce data security.

If your "replace from disk" was faulted or offline, it cannot be updated.

I am not sure about the behavior on a "too many error" situation.
I would then expect a "not updated= from disk faulted"
 
OK good, because a third drive on the vdev is showing as "resilvering". Is there a way to update that view so I can see if it's still resilvering ? As I mentioned earlier I found out rebooting the server did the trick, but that's not a great option.
 
I'm probably one of the last of the mohicans still running on OI with napp-it.. :eek:
Had some free time on my hands and after moving to esxi 5.5 u2 patch 3 (esxi 5.5.0,2143827) quite some time ago the next thing on my list is my favorite zfs system.

Downloaded Gea's last appliance nappit_15b and ran it for a couple of days. (using 2 vmxnet3 adapters)
I noticed that when trying to graceful shutdown or reboot the appliance, the vm would kernel panic often:
wZvQO7c.png


This also happens the first time when starting the vm (after I've used the power, graceful shutdown power button)
Code: 
root@vnastest02:/root# fmdump -Vp -u 78e5515c-ad62-4d89-c18c-92d7298a1e34
TIME                           UUID                                 SUNW-MSG-ID
Jul 26 2015 20:15:06.788389000 78e5515c-ad62-4d89-c18c-92d7298a1e34 SUNOS-8000-KL

  TIME                 CLASS                                 ENA
  Jul 26 20:15:05.8279 ireport.os.sunos.panic.dump_available 0x0000000000000000
  Jul 26 20:15:03.5497 ireport.os.sunos.panic.dump_pending_on_device 0x0000000000000000

nvlist version: 0
        version = 0x0
        class = list.suspect
        uuid = 78e5515c-ad62-4d89-c18c-92d7298a1e34
        code = SUNOS-8000-KL
        diag-time = 1437941706 785707
        de = fmd:///module/software-diagnosis
        fault-list-sz = 0x1
        fault-list = (array of embedded nvlists)
        (start fault-list[0])
        nvlist version: 0
                version = 0x0
                class = defect.sunos.kernel.panic
                certainty = 0x64
                asru = sw:///:path=/var/crash/unknown/.78e5515c-ad62-4d89-c18c-92d7298a1e34
                resource = sw:///:path=/var/crash/unknown/.78e5515c-ad62-4d89-c18c-92d7298a1e34
                savecore-succcess = 1
                dump-dir = /var/crash/unknown
                dump-files = vmdump.0
                os-instance-uuid = 78e5515c-ad62-4d89-c18c-92d7298a1e34
                panicstr = mutex_enter: bad mutex, lp=ffffffffc023ae98 owner=ffffff0192a5b780 thread=ffffff00073d1c40
                panicstack = unix:mutex_panic+58 () | unix:mutex_vector_enter+347 () | genunix:cv_timedwait_hires+107 () | genunix:cv_timedwait_sig_hires+1eb () | genunix:cv_timedwait_sig+49 () | fffffffff853be05 () | unix:thread_start+8 () |
                crashtime = 1437846388
                panic-time = Sat Jul 25 17:46:28 2015 UTC
        (end fault-list[0])

        fault-status = 0x1
        severity = Major
        __ttl = 0x1
        __tod = 0x55b53fca 0x2efddc88
Since my machine is pretty solid (still running OI without issue) I looked around the forums and found this old thread which exactly describes the same issue 2 years ago!:
http://hardforum.com/showthread.php?t=1783445

Just wanted to get a taste with other people's experience...is this issue still hanging around OmniOS ?

I've tried the workaround by editing the poweroff-vm-default and it seems to work. (esxi throws a error but there is no kernel panic)

If this bug is still lingering is that also the case with adding the 2nd cpu which cause high CPU usage as described in the thread ?:
http://hardforum.com/showthread.php?t=1783445
 
Last edited:
This was mainly a bug with ESXi 5.5 (initial release) and e1000.
This was fixed in 5.5U1 but with a new NFS bug introduced.

5.5U2 is quite solid but ESXi, especially with vmxnet3 is very picky
on some hardware. If the error on powerdown is the only remaining problen,
I would not care.

Other option
- use e1000 (mostly slower with a higher cpu load)
- try tools from ESXi 5.5 (tools in the 15b image are for ESXi 6 but they should work with 5.5)


beside that
OI 151a9 dev is a quite old illumos
OmniOS is quite up to date with regular updates and fixes.

Mostly OmniOS is more stable and faster.
 
_Gea said:
Its in the repo:
http://pkg.omniti.com/omniti-perl/en/index.shtml
other repos: http://omnios.omniti.com/wiki.php/Packaging

read:
http://omnios.omniti.com/wiki.php/GeneralAdministration#ConfigurePublishers

Configure Publishers #
List configured publishers: pkg publisher
Add a publisher (example): pkg set-publisher -g http://pkg.omniti.com/omniti-ms/ ms.omniti.com
Remove a publisher: pkg unset-publisher ms.omniti.com
Click to expand...

I've been trying to get TLS working for me to no avail. I've been trying on and off for months now. I even though outgoing.verizon.net with my verizon fios account info I could do port 25 smtp but for whatever reason that fails when I test it (with regular mail enabled in napp-it). That port works fine for my UPS events notifier, so I've no idea why napp-it won't take it.

Anyhow, I cannot add the repos Gea mentioned. I get the error quoted above or I get this:

Unable to contact valid package repository: http://pkg.omniti.com/omniti-perl-en-index.shtml/
Encountered the following error(s):
Transport errors encountered when trying to contact repository.
Reported the following errors:
http protocol error: code: 404 reason: Not Found
URL: 'http://pkg.omniti.com/omniti-perl-en-index.shtml'. (happened 4 times)
Click to expand...

When I list the repos I have, I get this:
PUBLISHER TYPE STATUS URI
omnios origin online http://pkg.omniti.com/omnios/r151012/
ms.omniti.com origin online http://pkg.omniti.com/omniti-ms/
root@Omni-Napp:/root#
Click to expand...

I've been able to do these steps:
perl -MCPAN -e shell
notest install Net::SSLeay
notest install IO::Socket::SSL
notest install Net::SMTP::TLS

I just am unable to run install pkg net-ssleay... when I do, the repos are checked and give me this:
The following pattern(s) did not match any allowable packages. Try
using a different matching pattern, or refreshing publisher information:
Click to expand...

I even tried this:
https://loadingsysadmin.wordpress.c...g-napp-it-to-send-email-alerts-through-gmail/
Step 6 gives me a could not fetch error.

Any help would be appreciated. Thanks.
 
For TLS email on OmniOS, you only need to enter as root on a console
(Answer questions with enter)

perl -MCPAN -e shell
notest install Net::SSLeay
notest install IO::Socket::SSL
notest install Net::SMTP::TLS
exit;

A manual install of net-ssleay via pkg install
is not needed as it is included in the CPAN setup.
http://napp-it.org/downloads/tls_en.html

If you are on OmniOS 151014, you need napp-it 0.9f6 dev
due a newer OpenSSL in this release
http://napp-it.org/downloads/changelog_en.html
 
_Gea said:
This was mainly a bug with ESXi 5.5 (initial release) and e1000.
This was fixed in 5.5U1 but with a new NFS bug introduced.

5.5U2 is quite solid but ESXi, especially with vmxnet3 is very picky
on some hardware. If the error on powerdown is the only remaining problen,
I would not care.

Other option
- use e1000 (mostly slower with a higher cpu load)
- try tools from ESXi 5.5 (tools in the 15b image are for ESXi 6 but they should work with 5.5)


beside that
OI 151a9 dev is a quite old illumos
OmniOS is quite up to date with regular updates and fixes.

Mostly OmniOS is more stable and faster.
Click to expand...

Thanks Gea,
So far haven't seen any issues other than what I've described..
Wasn't aware that the ESXI tools were for version 6 as the vm is named:
napp-it_15b_vm_for_ESXi_5.5u2-6.0.zip (May 28,2015) so I assumed it was for 5.5 and >

Any reason you give your Napp-IT Appliance 1 vcpu instead of 2?

Well I'll have a go with OmniOS in the next couple of days, (created the same local users as in my OI box), made sure that all Napp-IT settings are the same between the two boxes, have to export the pools, release my M1015 from OI and pass it to OmniOS and away we go..

if things fail I can always jump to ESXI 6..
 
Last edited:
_Gea said:
For TLS email on OmniOS, you only need to enter as root on a console
(Answer questions with enter)

perl -MCPAN -e shell
notest install Net::SSLeay
notest install IO::Socket::SSL
notest install Net::SMTP::TLS
exit;

A manual install of net-ssleay via pkg install
is not needed as it is included in the CPAN setup.
http://napp-it.org/downloads/tls_en.html

If you are on OmniOS 151014, you need napp-it 0.9f6 dev
due a newer OpenSSL in this release
http://napp-it.org/downloads/changelog_en.html
Click to expand...

I'd ran the MCPAN commands via putty, but still the smtp test fails.
I'm on OmniOS v11 151012.
I decided to update napp-it from 0.9f3 to f5 via the web interface and the webserver is down. Fixed it with wget.

So now I'm on OmniOS v11 151012 and napp-it 0.9f5. I reran the commands above as you mentioned and it says everything is already up to date. when I try a test smtp, it still fails. =/
I've tried both with proper authentication:
smtp.gmail.com
outgoing.verizon.net

Neither work.
I'm wondering if I'm having a network problem, since I also have issues adding repositories. I tried pinging out and having trouble...
 
Last edited:
nostradamus99 said:
Wasn't aware that the ESXI tools were for version 6 as the vm is named:
napp-it_15b_vm_for_ESXi_5.5u2-6.0.zip (May 28,2015) so I assumed it was for 5.5 and >

Any reason you give your Napp-IT Appliance 1 vcpu instead of 2?

Well I'll have a go with OmniOS in the next couple of days, (created the same local users as in my OI box), made sure that all Napp-IT settings are the same between the two boxes, have to export the pools, release my M1015 from OI and pass it to OmniOS and away we go..

if things fail I can always jump to ESXI 6..
Click to expand...

- ESXi tools are compatible to former versions of ESXi
- one vcpu is the default setting. You can modify if you want.
 
g0dM@n said:
I'd ran the MCPAN commands via putty, but still the smtp test fails.
I'm on OmniOS v11 151012.
I decided to update napp-it from 0.9f3 to f5 via the web interface and the webserver is down. Fixed it with wget.

So now I'm on OmniOS v11 151012 and napp-it 0.9f5. I reran the commands above as you mentioned and it says everything is already up to date. when I try a test smtp, it still fails. =/
I've tried both with proper authentication:
smtp.gmail.com
outgoing.verizon.net

Neither work.
I'm wondering if I'm having a network problem, since I also have issues adding repositories. I tried pinging out and having trouble...
Click to expand...


You should check your network settings first in menu system (manual ip, subnetmask, gateway, DNS or use DHCP for auto-config).
A "ping 8.8.8.8" to Google DNS must give an alive as well as a "ping www.google.com"

If this works but TLS not, you may need to update to OmniOS 151014 and napp-it 0.9f6 dev as
this takes care about some critical OpenSSL fixes of last year that may be needed for your mailprovider

0.9f6 is the current developer release If you are on the stable napp-it free, you can use the following key
to update to 0.9f6dev (copy/paste the next line to extension-register)
pro tls-test - 29.07.2015::mVqmVqsmVKTTstqmPqsmVKTTDmDs
 
Last edited:
I'm still using OI, mainly because I've installed wine and several windows apps I need. I spent quite some time to do this, as I'm not too experienced with Linux, much less Solaris.

Anyone using wine on OmniOS ?

_Gea said:
It is obvious that if your "replace from" disk was online, that it got updated as a failure of the ""replace to" disk will otherwise reduce data security.
Click to expand...

The fact that the pool stays in an "ONLINE" state and not "DEGRADED" is a good indication, I should have realized.
 
_Gea said:
You should check your network settings first in menu system (manual ip, subnetmask, gateway, DNS or use DHCP for auto-config).
A "ping 8.8.8.8" to Google DNS must give an alive as well as a "ping www.google.com"

If this works but TLS not, you may need to update to OmniOS 151014 and napp-it 0.9f6 dev as
this takes care about some critical OpenSSL fixes of last year that may be needed for your mailprovider

0.9f6 is the current developer release If you are on the stable napp-it free, you can use the following key
to update to 0.9f6dev (copy/paste the next line to extension-register)
pro tls-test - 29.07.2015::mVqmVqsmVKTTstqmPqsmVKTTDmDs
Click to expand...

I was able to ping primary DNS servers and google. I think for a brief moment yesterday I couldn't ping google. Anyway, I did update OmniOS to 151014. I hope I did it correctly as I'm still a beginner. OmniOS loads up, but the webserver is now down. I'm going to install 0.9f5 with wget again and once I get the webserver up, I'll try to put in your key, as I don't know how to do it via CLI.

Because of all the work I've been doing, I have a ton of boot environments. At what point should I clean that up (and how)? Thanks in advance!

*EDIT*
OK, I'm def having network issues. I can't even update via wget. I can't ping 8.8.8.8. Need to figure out how to fix network config via CLI... =/
 
Last edited:
_Gea said:
You can delete BEs in napp-it, menu snapshots
Prior an update of OmnisOS to 151014, read http://omnios.omniti.com/wiki.php/Upgrade_to_r151014

Update of napp-it:
enter the key in "extension >> register" and update via menu "about > update: download 0.9f6"
Click to expand...

Thanks. I'm hoping I can do 0.9f6 and then upgrade OmniOS.

I'm going back to 151012. I can't even use wget on 151014 because I'm unable to resolve DNS. I couldn't find a way to change the DNS server, just default gateway. I think the bulk of my problems is my network issues.

I have two NICS on the VM, one for NFS on 10.0.69.x/24, and one for LAN 192.168.69.x/24. I must have screwed something up in the past and clueless on how to fix it.
 
_Gea said:
The DNS entry (/etc/resolv.conf) can be edited in napp-it menu System > Network Eth >> DNS to something like (Google DNS)
nameserver 8.8.8.8

Maybe you need to add a default route, see
http://napp-it.org/downloads/omnios_en.html chapter 6.0
Click to expand...

Here are my settings and you can see how it fails 8.8.8.8 and eventually works on its own:


I rebooted, still no go, so I followed through adding the default route (which I'd done previously, as you can see):


Could this be because I have multiple phys adapters? I actually wanted one for management, one for LAN, and one for NFS for VMs to attach to. I left the management/LAN on the 192.168.69.x/24 subnet, and NFS on 10.0.69.x/24 network.
 
Multiple physical adapters with a dedicated management network to restrict access based on source/target ip and a firewall is the default professional setup. You only need to care that you normally have only one default gateway/route on one adapter.

Your setup seems quite ok but a ping to 8.8.8.8 does not work on first try, then it works and it does not on a reboot.

I would delete all links (either all beside management in menu system - network Eth under state or delete all at CLI like ipadm delete-if e1000g0)

and care about your default route (maybe the problem is there)
https://docs.oracle.com/cd/E36784_01/html/E37475/gmyag.html

delete (route flush) and recreate a default route
 
_Gea, I'm trying to download the vmware VM on your website but it doesn't work, in fact I can't download anything, I've tried from work and from home, is your website OK ? If so, is there a torrent I could use ?
 
Aesma said:
_Gea, I'm trying to download the vmware VM on your website but it doesn't work, in fact I can't download anything, I've tried from work and from home, is your website OK ? If so, is there a torrent I could use ?
Click to expand...

Yes, webserver lighttpd that serves the ESXi images via http seems down.
As it is 9:30 PM in Germany, I will fix this tomorrow morning.
 
Last edited:
Aesma said:
_Gea, I'm trying to download the vmware VM on your website but it doesn't work, in fact I can't download anything, I've tried from work and from home, is your website OK ? If so, is there a torrent I could use ?
Click to expand...

What ISO or EXE exactly are you looking for? I may be able to host it for you if I have it.
 
I'm in France so it's late too, don't worry I can wait !

g0dM@n : thanks but I can wait a bit.
 
Dear Gea,

I'm experiencing the pool timeout issue:

# dmesg
......
scsi: [ID 107833 kern.notice] /pci@0,0/pci15ad,7a0@18,1/pci15d9,691@0 (mpt_sas4):
Timeout of 60 seconds expired with 1 commands on target 25 lun 0.
scsi: [ID 107833 kern.warning] WARNING: /pci@0,0/pci15ad,7a0@18,1/pci15d9,691@0 (mpt_sas4):
Disconnected command timeout for target 25 w5000c50057758511.
.......

Motherboard: supermicro X9DR7-JLN4F
Chassis: SC826BE26-R920LPB, dual expander
HBA: aoc-s2308l-l8e
Disk: 12x seagate sas constellation ST2000NM0001
Omni sees 2 targets per disk.

----------

1 disk failure:
c9t5000C50057758513d0 Soft Errors: 0 Hard Errors: 23 Transport Errors: 229
Media Error: 8 Device Not Ready: 0 No Device: 12 Recoverable: 0

But Omni v151012 does not kick it out. It's just retry, and the whole pool is locked

I even set the command timeout to 5s:
http://everycity.co.uk/alasdair/2011/05/adjusting-drive-timeouts-with-mdb-on-solaris-or-openindiana/
But that disk still stay in the pool.

------

I must login shell, type zpool detach pool diskid to kick it out. (Nappit GUI is also timeout, because 'format' command is running on background).

Do you have any recommendation? This is the 3rd times i encounter this disk failure issue.
Best regards.
 
Last edited:
The problem is, that the disk gives transport and media errors only.
They basically work but are extremely slow.

The best recommendation is
- replace the affected disks and do some low level tests to repair them or mark them as bad

then
- try the newer OmniOS 151014 LTS
- try another firmware on your LSI
If you use the P20 firmware on your LSI 2308 (known to have problems), try an older one like P19.
 
Dear Gea,

151014 is un-usable with esxi 5.1U3. Esxi can mount the NFS share, but after some data read/write, that pool is not responding.
Even there is no I/O, i can't upload an ISO to it; browsing is ok, but uploading never finish.

Current LSI 2308 firmware is P16. Which version do you think 'best'?

So, it seems there is no solution for this kinds of timeout, right?

Best regards.
 
dualathlon said:
Dear Gea,

151014 is un-usable with esxi 5.1U3. Esxi can mount the NFS share, but after some data read/write, that pool is not responding.
Even there is no I/O, i can't upload an ISO to it; browsing is ok, but uploading never finish.

Current LSI 2308 firmware is P16. Which version do you think 'best'?

So, it seems there is no solution for this kinds of timeout, right?

Best regards.
Click to expand...

I cannot comment 5.1 U3, what I am using is 5.5U2 with 151014 which is very stable.
ESXi 5.5 initial is buggy regarding e1000 and ESXi 5.5 U1 is buggy regarding NFS.

I have not heard of problems with P16..P19 firmware. Only current P20 seems to be a problem.

ps
If 151012 worked, you may switch back to the last BE with this version.
You may use a testsetup to check a newer ESXi and a newer OmniOS with your hardware.
 
What are some options to synchronize UIDs and GIDs across multiple servers (Windows and Linux/Debian) for use with napp-it?

I'm only using this setup in a home network, but I have different user accounts for each service (media server, file server, etc) and I'd also like my own user account to be automatically replicated when I create a new server.

Would Active Directory be complete overkill, or what are some of the other geeks here using?

EDIT: Can Oracle Solaris become a domain controller?
 
Last edited:
1.
When you create a new user, you can can force a UID/GID on any Linux/Unix.
For a small home environment it may be the best to use the same UID/GID/PW on any server

2.
Active Directory (or LDAP) is a centralized user database and directory service
if you need the same user accounts on a group of machines.

3.
Not any server service can use such a directory service.
Mostly you are restricted to SMB.

3.
Solaris/ Omnios includes the software to easily become a member of an Active Directory domain.
If you want to setup a domain controller, you need Windows Server or SAMBA (on any Linux/Unix).
On Solaris, you can use SAMBA as well. In my own setups I use Windows Server and avoid SAMBA
as I prefer Solaris CIFS as an SMB clients for many reasons.
 
About stability, which i like to share:

I tested in the past OmniOS 151012 and VMWare ESXi 6, when it was out and it was NOT stable with NFS datastores. Then i switched back to ESXi 5.5U2 with the latest patch level and all was fine. Meanwhile i switched to OmniOS 151014 - 170cea2 from April, everything also fine with NFS Datastores. Last week i switched to the latest Patch Level ESXi 6.0.0b - ESXi-6.0.0-20150704001-standard (Build 2809209), NFS also working very stable, there was a bug fix regarding NFS in KB2111983:

"Slow NFS storage performance is observed on virtual machines running on VSA provisioned NFS storage. Delayed acknowledgements from the ESXi host for the NFS Read responses might cause this performance issue.

This patch resolves this issue by disabling delayed acknowledgements for NFS connections."

After a while everything was fine, i tested the latest OmiOS build 151014 - 7648372 from 27 July, and i was loosing again connection to NFS datastores. So this combination is NOT stable.

Since EXSi 5.1 i had various issues regarding NFS in combinations with OmniOS and ESXi, its a bit annoying to find out, what is working stable and what not. With latest ESXi 6 and OmniOS 151014 from
april, all is good for me.
 
Last edited:
@schleicher
I've been having the EXACT same issues as you had when using 151014 - 7648372 using ESXI 5.5u2 patch 3 (and up to the latest 5.5 patch as of August 2015)

Thanks Gea for bringing this to my attention! (saves me a lot of heart and headache!)
Will try again with your original appliance from April !
 
Last edited:
I've used napp-it and joined my Active Directory (using the Services -> SMB -> Active Directory), but do I need to configure anything else so that Solaris itself is aware of the domain and so I can login (ssh, ftp, etc) using my AD accounts?
 
Joining an AD is not a system wide join. It works only for the Solaris CIFS server to use the AD users with their Windows security id's.The CIFS servers store them as extended ZFS attributes so Solaris can act similar to an NTFS filesystem regarding permissions.

Other services like Apache or Proftp must care about a directory service for their own.
 
schleicher said:
About stability, which i like to share:

I tested in the past OmniOS 151012 and VMWare ESXi 6, when it was out and it was NOT stable with NFS datastores. Then i switched back to ESXi 5.5U2 with the latest patch level and all was fine. Meanwhile i switched to OmniOS 151014 - 170cea2 from April, everything also fine with NFS Datastores. Last week i switched to the latest Patch Level ESXi 6.0.0b - ESXi-6.0.0-20150704001-standard (Build 2809209), NFS also working very stable, there was a bug fix regarding NFS in KB2111983:

"Slow NFS storage performance is observed on virtual machines running on VSA provisioned NFS storage. Delayed acknowledgements from the ESXi host for the NFS Read responses might cause this performance issue.

This patch resolves this issue by disabling delayed acknowledgements for NFS connections."

After a while everything was fine, i tested the latest OmiOS build 151014 - 7648372 from 27 July, and i was loosing again connection to NFS datastores. So this combination is NOT stable.

Since EXSi 5.1 i had various issues regarding NFS in combinations with OmniOS and ESXi, its a bit annoying to find out, what is working stable and what not. With latest ESXi 6 and OmniOS 151014 from
april, all is good for me.
Click to expand...

Very interesting. I updated to ESXi 6 build 2809209 and OmniOS 151014 build 7648372 just a few days ago and haven't noticed any issues with NFS datastores at all.
 
dualathlon said:
Dear Gea,

I'm experiencing the pool timeout issue:

# dmesg
......
scsi: [ID 107833 kern.notice] /pci@0,0/pci15ad,7a0@18,1/pci15d9,691@0 (mpt_sas4):
Timeout of 60 seconds expired with 1 commands on target 25 lun 0.
scsi: [ID 107833 kern.warning] WARNING: /pci@0,0/pci15ad,7a0@18,1/pci15d9,691@0 (mpt_sas4):
Disconnected command timeout for target 25 w5000c50057758511.
.......

Motherboard: supermicro X9DR7-JLN4F
Chassis: SC826BE26-R920LPB, dual expander
HBA: aoc-s2308l-l8e
Disk: 12x seagate sas constellation ST2000NM0001
Omni sees 2 targets per disk.

----------

1 disk failure:
c9t5000C50057758513d0 Soft Errors: 0 Hard Errors: 23 Transport Errors: 229
Media Error: 8 Device Not Ready: 0 No Device: 12 Recoverable: 0

But Omni v151012 does not kick it out. It's just retry, and the whole pool is locked

I even set the command timeout to 5s:
http://everycity.co.uk/alasdair/2011/05/adjusting-drive-timeouts-with-mdb-on-solaris-or-openindiana/
But that disk still stay in the pool.

------

I must login shell, type zpool detach pool diskid to kick it out. (Nappit GUI is also timeout, because 'format' command is running on background).

Do you have any recommendation? This is the 3rd times i encounter this disk failure issue.
Best regards.
Click to expand...


As an update
There are some discussions/ work on timeout problems of the mpt sas driver
see illumos developer mailing list.

http://search.gmane.org/?query=6119...and&xP=Zmptsas&xFILTERS=Gos.illumos.devel---A
 
RyC said:
Very interesting. I updated to ESXi 6 build 2809209 and OmniOS 151014 build 7648372 just a few days ago and haven't noticed any issues with NFS datastores at all.
Click to expand...

Probably it depends on the hardware you are using. I can only report from my experience...
 
You must log in or register to reply here.
Back
Top