unresponsive system during moves

16 views
Skip to first unread message

vajonam

unread,
Aug 14, 2019, 7:05:23 PM8/14/19
to EON ZFS Storage
I completed the upgrade to the latest and updated the pools and file systems.  since I do need to move to more recent code. all went well with the upgrade. 

I do notice that sometimes when I "move" data between the drives the whole system becomes unresponsive and I can't even ping the box from the outside.  Are there anything I need to tune to for the system to remain responsive? and for other operations to continue? 

Andre Lue

unread,
Aug 14, 2019, 10:35:48 PM8/14/19
to EON ZFS Storage on behalf of vajonam
No, you shouldn't have to tune anything. Does the system recover from the freeze or do you have to power it off/on?

Is this on a VM or a bare metal system?

On Wed, Aug 14, 2019 at 7:05 PM vajonam via EON ZFS Storage <eonstorage+APn2wQeaGveeVtAFGo6St...@googlegroups.com> wrote:
I completed the upgrade to the latest and updated the pools and file systems.  since I do need to move to more recent code. all went well with the upgrade. 

I do notice that sometimes when I "move" data between the drives the whole system becomes unresponsive and I can't even ping the box from the outside.  Are there anything I need to tune to for the system to remain responsive? and for other operations to continue? 

--
You received this message because you are subscribed to the Google Groups "EON ZFS Storage" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/eonstorage/5610c6d6-b99e-486e-92f8-875f08eabf3c%40googlegroups.com.

Manojav Sridhar

unread,
Aug 14, 2019, 11:33:42 PM8/14/19
to EON ZFS Storage on behalf of dre2kse
The system recovers from the freeze. sometimes takes like 30 seconds. 

its a bare metal system. no VMs, I am not even running anything on the box just serving up files. 



Manojav Sridhar

unread,
Aug 14, 2019, 11:39:43 PM8/14/19
to EON ZFS Storage on behalf of dre2kse
hmm this is bit concerning looks I am not able to very much on that box in terms of writes.. before it locks up the system. 

This is what I see in the client logs. 

[  537.443278] nfs: server getafix not responding, timed out
[  545.639689] nfs: server getafix not responding, timed out

vajonam

unread,
Aug 14, 2019, 11:43:23 PM8/14/19
to EON ZFS Storage
what rsize and wsize should I be using the nfs mounts on the clients? 



On Wednesday, August 14, 2019 at 11:39:43 PM UTC-4, vajonam wrote:
hmm this is bit concerning looks I am not able to very much on that box in terms of writes.. before it locks up the system. 

This is what I see in the client logs. 

[  537.443278] nfs: server getafix not responding, timed out
[  545.639689] nfs: server getafix not responding, timed out

On Wed, Aug 14, 2019 at 11:33 PM Manojav Sridhar <man...@manojav.com> wrote:
The system recovers from the freeze. sometimes takes like 30 seconds. 

its a bare metal system. no VMs, I am not even running anything on the box just serving up files. 



On Wed, Aug 14, 2019 at 10:35 PM EON ZFS Storage on behalf of dre2kse <eonst...@googlegroups.com> wrote:
No, you shouldn't have to tune anything. Does the system recover from the freeze or do you have to power it off/on?

Is this on a VM or a bare metal system?

On Wed, Aug 14, 2019 at 7:05 PM vajonam via EON ZFS Storage <eonstorage+APn2wQeaGveeVtAFGo6StkzOM-yyat2Qk_yK0TpsH4zA9NaHLqHYt@googlegroups.com> wrote:
I completed the upgrade to the latest and updated the pools and file systems.  since I do need to move to more recent code. all went well with the upgrade. 

I do notice that sometimes when I "move" data between the drives the whole system becomes unresponsive and I can't even ping the box from the outside.  Are there anything I need to tune to for the system to remain responsive? and for other operations to continue? 

--
You received this message because you are subscribed to the Google Groups "EON ZFS Storage" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+unsubscribe@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "EON ZFS Storage" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+unsubscribe@googlegroups.com.
Message has been deleted

vajonam

unread,
Aug 15, 2019, 12:01:02 AM8/15/19
to EON ZFS Storage
Not sure what to do... dead in the water.. i ran it for 2 days without doing the zfs / zpool upgrade and no issues... crap.. 

vajonam

unread,
Aug 15, 2019, 12:08:24 AM8/15/19
to EON ZFS Storage
even the console is kind of busy, the return key moves the screen up, but I dont see the login prompt.. this is the longest its frozen.. maybe worth a reboot

vajonam

unread,
Aug 15, 2019, 12:11:13 AM8/15/19
to EON ZFS Storage
and its back.

on the bootup screen I see an error about no mircocode for my intel processor is this a concern? 

last pid:  5560;  load avg:  76.5,  47.9,  31.2;       up 0+16:54:23                                                    04:10:16
35 processes: 33 sleeping, 1 stopped, 1 on cpu
CPU states: 99.9% idle,  0.5% user,  0.1% kernel,  0.0% iowait,  0.0% swap
Memory: 4094M phys mem, 501M free mem, 4096M total swap, 4091M free swap

   PID USERNAME LWP PRI NICE  SIZE   RES STATE    TIME    CPU COMMAND
  5547 root       1  59    0 4820K 2844K cpu/1    0:00  0.01% top
  1190 root      15  59    0   11M 4996K sleep    0:01  0.00% smbd
  1503 root       1  59    0 6504K 3812K sleep    0:01  0.00% ntpd
  5502 root       1  59    0 8252K 4904K sleep    0:00  0.00% sshd
     8 root      11  59    0 6068K 2184K sleep    0:00  0.00% svc.startd
    10 root      14  59    0 7104K 4284K sleep    0:03  0.00% svc.configd
   476 daemon     6  59    0 4680K 2792K sleep    0:00  0.00% idmapd
  1134 root       1  59    0 4140K   36K sleep    0:00  0.00% ipmon
   481 root       7  59    0 4820K 1888K sleep    0:00  0.00% devfsadm
  1322 root       1  59    0 1624K  876K sleep    0:00  0.00% utmpd
  2443 daemon     2  60  -20 3032K 1768K sleep   88:31  0.00% nfsd
  5522 root       1  59    0 2500K 1892K stop     0:00  0.00% iostat
  5542 root       1  59    0 8252K 4908K sleep    0:00  0.00% sshd
  5541 root       1  59    0 6520K 3344K sleep    0:00  0.00% sshd
  5501 root       1  59    0 6520K 3312K sleep    0:00  0.00% sshd





On Thursday, August 15, 2019 at 12:01:02 AM UTC-4, vajonam wrote:

Andre Lue

unread,
Aug 15, 2019, 12:29:55 AM8/15/19
to EON ZFS Storage on behalf of vajonam
I think this is more evident in the older versions. I will try to get you a newer release or a newer version to test, depending on time BW.

vajonam

unread,
Aug 15, 2019, 12:40:28 AM8/15/19
to EON ZFS Storage
Thank you so much! Even though am using this for personal use, I only think its fair that I donate to your cause this has given me really good performance over the years ! 


On Thursday, August 15, 2019 at 12:29:55 AM UTC-4, dre2kse wrote:
I think this is more evident in the older versions. I will try to get you a newer release or a newer version to test, depending on time BW.

On Wed, Aug 14, 2019 at 11:33 PM EON ZFS Storage on behalf of vajonam <eonst...@googlegroups.com> wrote:
The system recovers from the freeze. sometimes takes like 30 seconds. 

its a bare metal system. no VMs, I am not even running anything on the box just serving up files. 



On Wed, Aug 14, 2019 at 10:35 PM EON ZFS Storage on behalf of dre2kse <eonst...@googlegroups.com> wrote:
No, you shouldn't have to tune anything. Does the system recover from the freeze or do you have to power it off/on?

Is this on a VM or a bare metal system?

On Wed, Aug 14, 2019 at 7:05 PM vajonam via EON ZFS Storage <eonstorage+APn2wQeaGveeVtAFGo6StkzOM-yyat2Qk_yK0TpsH4zA9NaHLqHYt@googlegroups.com> wrote:
I completed the upgrade to the latest and updated the pools and file systems.  since I do need to move to more recent code. all went well with the upgrade. 

I do notice that sometimes when I "move" data between the drives the whole system becomes unresponsive and I can't even ping the box from the outside.  Are there anything I need to tune to for the system to remain responsive? and for other operations to continue? 

--
You received this message because you are subscribed to the Google Groups "EON ZFS Storage" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eonst...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "EON ZFS Storage" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eonst...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "EON ZFS Storage" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eonst...@googlegroups.com.

vajonam

unread,
Aug 15, 2019, 12:54:40 AM8/15/19
to EON ZFS Storage
Also for more info. I am using ZIL log device to speed up writes maybe this is what is causing this runaway jobs? 


                              capacity     operations    bandwidth
pool                       alloc   free   read  write   read  write
-------------------------  -----  -----  -----  -----  -----  -----
gobi                       2.78T   868G     49     52   199K  5.24M
  mirror                   2.78T   868G     49     52   199K  5.24M
    c0t50014EE2B94B1268d0      -      -     20     52  83.0K  5.24M
    c0t50014EE2B94B2234d0      -      -     28     52   115K  5.24M
logs                           -      -      -      -      -      -
  c0t5000000000000000d0    37.8M   222G      0      0      0      0
-------------------------  -----  -----  -----  -----  -----  -----

vajonam

unread,
Aug 15, 2019, 1:39:28 AM8/15/19
to EON ZFS Storage
Also note that if set the "sync" on my nfs clients this seems to throttle the transfer into the ZFS file system and doesn't become unresponsive. 


On Thursday, August 15, 2019 at 12:29:55 AM UTC-4, dre2kse wrote:
I think this is more evident in the older versions. I will try to get you a newer release or a newer version to test, depending on time BW.

On Wed, Aug 14, 2019 at 11:33 PM EON ZFS Storage on behalf of vajonam <eonst...@googlegroups.com> wrote:
The system recovers from the freeze. sometimes takes like 30 seconds. 

its a bare metal system. no VMs, I am not even running anything on the box just serving up files. 



On Wed, Aug 14, 2019 at 10:35 PM EON ZFS Storage on behalf of dre2kse <eonst...@googlegroups.com> wrote:
No, you shouldn't have to tune anything. Does the system recover from the freeze or do you have to power it off/on?

Is this on a VM or a bare metal system?

On Wed, Aug 14, 2019 at 7:05 PM vajonam via EON ZFS Storage <eonstorage+APn2wQeaGveeVtAFGo6StkzOM-yyat2Qk_yK0TpsH4zA9NaHLqHYt@googlegroups.com> wrote:
I completed the upgrade to the latest and updated the pools and file systems.  since I do need to move to more recent code. all went well with the upgrade. 

I do notice that sometimes when I "move" data between the drives the whole system becomes unresponsive and I can't even ping the box from the outside.  Are there anything I need to tune to for the system to remain responsive? and for other operations to continue? 

--
You received this message because you are subscribed to the Google Groups "EON ZFS Storage" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eonst...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "EON ZFS Storage" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eonst...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "EON ZFS Storage" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eonst...@googlegroups.com.

Andre Lue

unread,
Aug 15, 2019, 8:44:00 AM8/15/19
to EON ZFS Storage on behalf of Donovan Kaardal
The data move between filesystems is over nfs only? Smb only? Or nfs, smb at the same time?

To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/eonstorage/3f5417bf-aa32-445a-9546-0f792acb69f7%40googlegroups.com.

vajonam

unread,
Aug 15, 2019, 8:43:21 PM8/15/19
to EON ZFS Storage
On NFS. I am using Ubuntu 18.04 clients over NFSv4. I have cifs enabled but don’t use much.

Andre Lue

unread,
Aug 15, 2019, 8:50:07 PM8/15/19
to EON ZFS Storage on behalf of Donovan Kaardal
If you're using nfsv4 on clients, make sure you configure eon nfs svr to match. I think its v3 by default which makes sense when you enable sync. Test v3 client, svr , v4 pair. I think v3 pair may perform better.

On Thu, Aug 15, 2019, 8:43 PM vajonam via EON ZFS Storage <eonstorage+APn2wQeaGveeVtAFGo6St...@googlegroups.com> wrote:
On NFS. I am using Ubuntu 18.04 clients over NFSv4. I have cifs enabled but don’t use much.

--
You received this message because you are subscribed to the Google Groups "EON ZFS Storage" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+...@googlegroups.com.

vajonam

unread,
Aug 16, 2019, 6:44:18 AM8/16/19
to EON ZFS Storage
Okay let me check the server versions and client versions.

vajonam

unread,
Aug 16, 2019, 6:47:29 AM8/16/19
to EON ZFS Storage
Just checked this, both sides are NFSv4 and the mounts are showing up with nfsv4
Message has been deleted

vajonam

unread,
Aug 16, 2019, 8:35:30 AM8/16/19
to EON ZFS Storage
tested out these combos. 


client -- server -- result
---------------------------
v3        v3        freeze
v4        v3        freeze
v4        v4        freeze
v3,udp    v3        freeze
v4,udp    v3        freeze
v4,udp    v4        freeze
v3,sync   v3        okay,but slower
v4,sync   v4        okay,but slower



vajonam

unread,
Aug 30, 2019, 10:14:34 AM8/30/19
to EON ZFS Storage
@dre2ske? any updates for me?

On somewhat related note, zpool import -a doesnt seem to setup the NFS shares.

I have to do zfs sharenfs pool/dataset  after for 5 of my 7 pools. Once I do this however all is well.

I have re-added and SSD removed it as ZIL device, and added it as a swap device. writes are overall slower, but less lockups. 
Reply all
Reply to author
Forward
0 new messages