SOLVED Truenas U1.1 fix schedule > waiting for U2 !

[tiberiusQ]

Thx & Greets ?

 

[Kris Moore]

We're looking at early next week, but waiting on a review for a couple performance related fixes still. Will known for sure in next day or so.

 

[tiberiusQ]

THX!

 

[headwhacker]

Is U1.1 on track for release this week?

 

[JesperA]

Yeah, i was also wondering if it was on track for release this week but then i saw this:

Reddit - The heart of the internet

www.reddit.com

https://jira.ixsystems.com/browse/NAS-108627

A silent corruption bug doesn't seem fun, so i have the SMB bug with macOS clients & potentially a silent corruption storage system, so, maybe IX would suggest us to go back to 11.3-U5 & wait for atleast 12.0-U2? Currently it doesn't seem to be safe running on any 12-train?

 

[andrema2]

Yeah, i was also wondering if it was on track for release this week but then i saw this:

Reddit - The heart of the internet

www.reddit.com

https://jira.ixsystems.com/browse/NAS-108627

A silent corruption bug doesn't seem fun, so i have the SMB bug with macOS clients & potentially a silent corruption storage system, so, maybe IX would suggest us to go back to 11.3-U5 & wait for atleast 12.0-U2? Currently it doesn't seem to be safe running on any 12-train?

Going back to 11.3-U5 after upgrade the pool, is it possible ? Or there were not upgrades on 12 ?

 

[Patrick M. Hausen]

Going back to 11.3-U5 after upgrade the pool, is it possible ?

After upgrade of the pool - no.

 

[awasb]

Since the silent data corruption seems to be exclusively linked with VMs under certain (rather special) circumstances and the arbitrary smb crashing with macOS clients just occurred with 12.0-U1, I switched back to 12.0-RELEASE and everything seems fine so far (the reporting bugs aside). Not much effort, just a reboot. "My" users are happy.

*coughing* "Move along, there is nothing to see here." *coughing*

 

[winnielinnie]

Concerning the "silent" data corruption, is it implied that not even a Scrub will detect corrupted blocks / data?

 

[tiberiusQ]

I do run 12U1 and beside the know issues I do experience a new strange behaviour-

All windows 10 clients seem to genarate insanely traffic while simply browsing the smb shares- Is this a known issue with this Truenas version or a new Windows feature ?

ref.

Truenas12x - Slow smb browsing !?

Dear all, With Truenas12U1 I do experience a very slow browsing performance and Windows 10 clients genarate insanely traffic while simply browsing the smb shares + The explorer crashed quiet often. Especially for vpn users it makes it nearly impossible to work ;-( The dataset for my share...

www.truenas.com

 

[awasb]

It's out and — as far as first stress tests show — the smb seg faults are seemed to be gone.

THX!

 

[awasb]

Too early. Still crashing ... three concurrent backups (approx 500GB each)

$ tail -f /var/log/samba4/log.smbd

[2021/01/16 16:44:27.197702, 0] ../../lib/util/fault.c:79(fault_report)


===============================================================


[2021/01/16 16:44:27.197795, 0] ../../lib/util/fault.c:80(fault_report)


INTERNAL ERROR: Signal 10 in pid 3270 (4.12.9)


If you are running a recent Samba version, and if you think this problem is not yet fixed in the latest versions, please consider reporting this bug, see htt


ps://wiki.samba.org/index.php/Bug_Reporting


[2021/01/16 16:44:27.197892, 0] ../../lib/util/fault.c:86(fault_report)


===============================================================


[2021/01/16 16:44:27.197954, 0] ../../source3/lib/util.c:830(smb_panic_s3)


PANIC (pid 3270): internal error


[2021/01/16 16:44:27.200466, 0] ../../lib/util/fault.c:265(log_stack_trace)


BACKTRACE: 6 stack frames:


#0 0x801952d17 <log_stack_trace+0x37> at /usr/local/lib/samba4/libsamba-util.so.0


#1 0x802ccfb86 <smb_panic_s3+0x56> at /usr/local/lib/samba4/libsmbconf.so.0


#2 0x801952b07 <smb_panic+0x17> at /usr/local/lib/samba4/libsamba-util.so.0


#3 0x801952eee <log_stack_trace+0x20e> at /usr/local/lib/samba4/libsamba-util.so.0


#4 0x801952ae9 <fault_setup+0x59> at /usr/local/lib/samba4/libsamba-util.so.0


#5 0x80fd48c20 <_pthread_sigmask+0x530> at /lib/libthr.so.3


[2021/01/16 16:44:27.200798, 0] ../../source3/lib/dumpcore.c:315(dump_core)


dumping core in /var/db/system/cores

Rolling back to TN12.0 RELEASE.

 

[djjaeger82]

Too early. Still crashing ... three concurrent backups (approx 500GB each)

Samba fix was not in u1.1, it was pushed to U2 since the data corruption fix was prioritized for U1.1. Be very careful running 12.0 or 12.0u1. For the data corruption severe issue I'd either roll back to 11.3 or keep 12.0u1.1 with the samba bug

 

[Yorick]

All windows 10 clients seem to genarate insanely traffic while simply browsing the smb shares

This happened to me just now after a fresh TrueNAS 12.0-U1.1 start, and Explorer crashed. After restarting Explorer, browsing was fast, and ARC use had grown to 20 GiB.

I'd say it has to do with fetching metadata. Give your TrueNAS enough RAM for ARC, that might just solve it. I had 32 GiB in this TrueNAS server and now increased it to 64 GiB. What are you running with?

 

[awasb]

@ djjaeger82: Tanks for the hint. Still ... I don't run any virtual instances on TrueNAS, so the data corruption bug in ZFS (as far as I know and understand it, that is) isn't an issue for me/us.

 

[tiberiusQ]

This happened to me just now after a fresh TrueNAS 12.0-U1.1 start, and Explorer crashed. After restarting Explorer, browsing was fast, and ARC use had grown to 20 GiB.

I'd say it has to do with fetching metadata. Give your TrueNAS enough RAM for ARC, that might just solve it. I had 32 GiB in this TrueNAS server and now increased it to 64 GiB. What are you running with?

Ahh - In this box I do also have 32gb ram, I may do uprade to 64 as well....

 

[gwaitsi]

Still hasn't solved the issue of:

Code:

nas 1 2021-01-17T14:09:19.315783+01:00 nas.net.lan mountd 1546 - - can't open /etc/zfs/exports
nas 1 2021-01-17T14:09:19.316993+01:00 nas.net.lan mountd 1546 - - can't open /etc/zfs/exports

These errors are triggered by the jails mounted drives and only started appearing after upgrading from 11.5 to 12.0

 

[CHL]

...silent data corruption seems to be exclusively linked with VMs under certain (rather special) circumstances...

is there some more information about this out there? how to know when I may run into this BEFORE it happens? data corruption on zfs was never worth a thought... up to now?

 

[Ericloewe]

Update to U1.1 and render that discussion a lot less relevant.

 

[tiberiusQ]

Yes, please somebody close that thread - I already taged it as solved.