Benni.blanko
Dabbler
- Joined
- Dec 18, 2021
- Messages
- 31
Hi there,
I'm trying to create a veracrypt container file on a Truenas Core 13.0-U5.2.
While the creation is running, my Windows 10 PC is writing with 500MBit/s to the device (over the 1GB Ethernet interface).
Also ALL other SMB connections are timing out/throwing errors on the Truenas box during this process. SMB seems to be completely unusable.
And this happens even on creating a small 20GB veracrypt container. Don't think a 3TB one ...
Is that a known behaviour? One connection can block everything else?
Isn't SMB here capable of handling multiple (even high loading) connections?
Should it even run faster than 500Mbit/s on SMB?
Replication operations to another Truenas Box are capable of fully utilizing the 1GB interface.
Truenas CPU-utilization is not alarming during this operation. All other services are also running fine (NFS, webgui, SSH).
top show:
last pid: 2754; load averages: 1.22, 0.94, 0.55 up 0+00:12:11 16:52:33
58 processes: 2 running, 56 sleeping
CPU: 8.8% user, 0.0% nice, 19.6% system, 0.0% interrupt, 71.5% idle
Mem: 1178M Active, 193M Inact, 1103M Wired, 29G Free
ARC: 266M Total, 72M MFU, 158M MRU, 11M Anon, 2973K Header, 17M Other
141M Compressed, 400M Uncompressed, 2.83:1 Ratio
Swap: 8192M Total, 8192M Free
PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND
2623 root 1 93 0 187M 161M CPU2 2 3:18 67.55% smbd
993 root 31 21 0 426M 287M kqread 1 0:45 0.38% python3.9
2754 root 1 21 0 14M 4160K CPU1 1 0:00 0.25% top
2740 root 1 20 0 14M 4168K select 3 0:00 0.07% top
1765 root 4 20 0 34M 6876K select 2 0:01 0.06% vmtoolsd
959 root 1 20 0 11M 2012K select 1 0:00 0.01% devd
2208 www 1 20 0 38M 11M kqread 0 0:00 0.01% nginx
2741 root 1 20 0 20M 9124K select 2 0:00 0.01% sshd
2046 root 1 20 0 28M 2880K select 2 0:00 0.01% mountd
2110 root 8 20 0 49M 13M select 2 0:01 0.01% rrdcached
2081 ntpd 1 20 0 17M 6884K select 0 0:00 0.01% ntpd
975 root 1 20 0 19M 6576K select 2 0:00 0.00% zfsd
2055 root 16 52 0 12M 2672K rpcsvc 1 0:00 0.00% nfsd
Configuration:
HPE DL380 with 10Core Xeon E5-2640V4 CPU (10C/20T 2,4Ghz/3,2Ghz) Total 144GB RAM, 1GB/s Ethernet port used
Truenas running as VM within ESXi 6.5.0U3, configured for 4vCPU (normally running fine with 2vCPU, increased to 4 now but no change in behaviour), 32GB RAM
LSI 2008 per PCI devicethru given to Truenas VM; several HDD and SDD attached
This setup runs well for >1,5 years now and I just fell into this problem while creating the veracrypt container.
Regards
BB
I'm trying to create a veracrypt container file on a Truenas Core 13.0-U5.2.
While the creation is running, my Windows 10 PC is writing with 500MBit/s to the device (over the 1GB Ethernet interface).
Also ALL other SMB connections are timing out/throwing errors on the Truenas box during this process. SMB seems to be completely unusable.
And this happens even on creating a small 20GB veracrypt container. Don't think a 3TB one ...
Is that a known behaviour? One connection can block everything else?
Isn't SMB here capable of handling multiple (even high loading) connections?
Should it even run faster than 500Mbit/s on SMB?
Replication operations to another Truenas Box are capable of fully utilizing the 1GB interface.
Truenas CPU-utilization is not alarming during this operation. All other services are also running fine (NFS, webgui, SSH).
top show:
last pid: 2754; load averages: 1.22, 0.94, 0.55 up 0+00:12:11 16:52:33
58 processes: 2 running, 56 sleeping
CPU: 8.8% user, 0.0% nice, 19.6% system, 0.0% interrupt, 71.5% idle
Mem: 1178M Active, 193M Inact, 1103M Wired, 29G Free
ARC: 266M Total, 72M MFU, 158M MRU, 11M Anon, 2973K Header, 17M Other
141M Compressed, 400M Uncompressed, 2.83:1 Ratio
Swap: 8192M Total, 8192M Free
PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND
2623 root 1 93 0 187M 161M CPU2 2 3:18 67.55% smbd
993 root 31 21 0 426M 287M kqread 1 0:45 0.38% python3.9
2754 root 1 21 0 14M 4160K CPU1 1 0:00 0.25% top
2740 root 1 20 0 14M 4168K select 3 0:00 0.07% top
1765 root 4 20 0 34M 6876K select 2 0:01 0.06% vmtoolsd
959 root 1 20 0 11M 2012K select 1 0:00 0.01% devd
2208 www 1 20 0 38M 11M kqread 0 0:00 0.01% nginx
2741 root 1 20 0 20M 9124K select 2 0:00 0.01% sshd
2046 root 1 20 0 28M 2880K select 2 0:00 0.01% mountd
2110 root 8 20 0 49M 13M select 2 0:01 0.01% rrdcached
2081 ntpd 1 20 0 17M 6884K select 0 0:00 0.01% ntpd
975 root 1 20 0 19M 6576K select 2 0:00 0.00% zfsd
2055 root 16 52 0 12M 2672K rpcsvc 1 0:00 0.00% nfsd
Configuration:
HPE DL380 with 10Core Xeon E5-2640V4 CPU (10C/20T 2,4Ghz/3,2Ghz) Total 144GB RAM, 1GB/s Ethernet port used
Truenas running as VM within ESXi 6.5.0U3, configured for 4vCPU (normally running fine with 2vCPU, increased to 4 now but no change in behaviour), 32GB RAM
LSI 2008 per PCI devicethru given to Truenas VM; several HDD and SDD attached
This setup runs well for >1,5 years now and I just fell into this problem while creating the veracrypt container.
Regards
BB