Can you do a crash test for me?

I think it is better to repeat all my story:

- bought readynas102 with two wd drives in compatibility list to store: some gigabytes of documents and for one (ONE) nfs share to use with pve/proxmox virtualizer. As a developer I have a virtual machine for each customer. So I have many vm but very light usage (one or two together). I chose readynas for: 1) snapshots 2) zfs like protection 3) debian based ( I hoped to do a plugin for ceph) 4) BECAUSE A CERTIFIED NAS SAVES ME A LOT OF TIME COMPARED TO A HOME MADE SOLUTION
- nas worked very well for one month then started crashing. I mean no reply from web interface, nfs protocol not responding and so
- worse when I powered up the nas again the raid1 rebuild was infinite. Here the nas seems to crash but it was only very slow: if I waited one day then I got some reply from the nas
- contacted the technical support that told me to do some hardware tests and then after some time told me to upgrade to new firmware just released that moved sectors from 4096 to bigger size.
- I had to reformat drives...
- so I bought a synology nas and faster sshd drives and moved my data on it
- synology has no snapshots so each day I do a full vm backup of around 200 gb to another disk. But guess what? Synology after few backups crashes too... if I ssh into it I see a load of 20 without doing anything
- so I take back readynas, put new sshd drives, resetted to factory defaults
- now I had the idea to do a bonnie++ test before putting my data on it
- test crashed so contacted technical support but they told me that my nas has no hardware problem and the warranty is out for software problems... if I can prove them that it is an hardware problem they will reply to me.......
- so I have written in this forum
- you suggested me to try less features
- I made an nfs share with only snapshots. Bonnie test passed.
- so I PUT AGAIN MY VMS ON IT
- guess what? now it is the third time that it crashes. And without IO LOAD AT ALL
- this time I enabled ssh
- TODAY i DISCOVERED THE NAS THAT WAS CRASHING BUT SSH WAS WORKING

So I logged in and I discovered:

- 0% cpu usage
- 100mb of free memory no swap usage at all
- no i/o (installed dstat to check)
- stopped apache2 but I was not able to restart it
- can do "ls" in most folders but not in "apps"
- SYSTEM LOAD OF 15!!!
- "sync" command never ends and does no i/o (checked with dstat)

Thanks again for interest,
Mario

Which apps do you have installed?

Can you update to 6.2.1-RC2 and try lowering the NFS thread count?

mdgm wrote:
Which apps do you have installed? Can you update to 6.2.1-RC2 and try lowering the NFS thread count?

Actually I resetted to factory defaults and I have NO apps installed.
Only 4 nfs shares: test1, test2,test3 for tests and one with 200gb of virtual machines.
I will try to update to 6.2.1 actually I have 8 threads that is minimum I can choose.
Nfs v4 disabled.

I have just installed 6.2.2 and problem persists.
Actually with fewer NFS threads the NAS hangs faster.
So if I:

- use 32 threads
- never do full backups (200gb of data to read)
- avoid stress tests

The NAS let me do my work without hanging.

It seems to me that all people that reply in this thread are sure that the problem is due to the fact that the nas is an entry level one and I should not abuse it.
But here we are talking about a nas that stops working and not about a nas that is too slow.
And I see that no one dared to try a bonnie++ test on his nas.
I also repeat that I connected with ssh to the nas and I have seen low cpu usage, half of the mem used, now swap at all used but a load of 12 without disk activity.
Has the nas been tested with long lasting nfs connections?

Thanks again,
Mario

After few weeks where I have disabled backups and so I do few mb of i/o in a day I confirm that nas hangs again.
So even not using it (but keeping nfs connections always on) crashes the nas.

So this confirm that it is a bug due to long standing nfs connections or something similar. Probably after some time a nfs thread stops working. So with 32 threads, low i/o two weeks are needed before all nfs stops responding.

But I see that:

1) netgear is not interested in bug reports but it is only worried to let me know that it do that warranty on software has expired
2) in the forum I see no one interested in try a bonnie++ on his nas to discover if it is flawed too
3) I have lost many weeks of my time that was precisely the main thing I want to avoid buying a commercial nas.

So now I will throw it in the trashcan and I build a serious nas based on freenas.

There are more recent threads on NFS issues and general lockups, not sure if you have checked them. Netgear seems to be investigating.

You could also potentially try 6.2.3 beta. viewtopic.php?f=154&t=78866

StephenB wrote:
There are more recent threads on NFS issues and general lockups, not sure if you have checked them. Netgear seems to be investigating.

I search the forum for "nfs" and I find nothing... where are them?

Thanks

There's a few general lockup threads - some include:
viewtopic.php?f=20&t=78697&hilit=lockup
viewtopic.php?f=65&t=79326&p=446237&hilit=lock#p446237

Here's one specific to NFS: viewtopic.php?f=7&t=78899&p=443277&hilit=lock#p443277

Can you send your logs?: http://readynas.com/kb/faq/misc/how_do_i_send_all_logs

The nas locks at least one time a day. Working with it is impossible.
I have already sent logs to technical support and they first told me that was some hardware issue then it was software issue, but only when software support was expired.
This is a sub par product with sub par support.
So I now will send logs only when someone will tell me that:
1) has tried the bonnie++ test
2) confirms that the nas hangs with nfs usage