AC68U - IO Error - SAMBA & USB3 - Multiple FW tested

[vapochilled]

Hello everyone,

We just got the Asus AC68U for a small office.
12 people, 3x connect by Lan, 7x by 5Ghz Wireless and 2x by 2.4Ghz.
We already tested 3 firmware's:
- stock 45xx firmware
- stock 4887 (last FW from 27 April)
- Merlin Beta 3 (the last one from March)


With all of them we experience IO corruption issues when we try to copy 50mb of files.
System env:
- WD elements 1TB - USB3 port
- NTFS
- Using access with 5 users.



The fails are random. I've seen multiple problems like this across the web.
People that have IO issues during the copies.
Some people even changed the filesystem to EXT4 but the problem persisted.


Do you have a solution for this? Otherwise I will just send this router back while we're on time.
I've tried some samba tweaks over the /etc/smb.conf file but still no go.
Are these problems related to buffer?
Permissions?
One thing I've noticed is that if you run a top or free command and if you keep your eye on the memory, this memory will decreased from the 256mb to 0mb during the copy.
So, if you copy 512Mb the RAM exhaust's.


Any ideas guys? We're running out of time and the people in the office are getting tired because all the time we have the IO issue for a shirtty 20mb copy (in this case it is not even exhausting the memory ):
- the router webpage becomes inaccessible
- the shares don't work anymore
- im not able to reach the network printer


What can we do?
- tweak Samba?
- change to usb 2 port?
- use the NFS sharing instead provided with Merlin FW?
- change to Alpha 4 and try again?
- take this s*** back?

 

[L&LD]

The version you want running is either the Asus 583 or the RMerlin version 374.40 alpha4.

Hopefully, one of those firmwares will fix your issues.

 

[vapochilled]

I will give them a try.
But all these issues seem to happen is most of the firmwares and I also don't understand that memory leak of consuming all the RAM during write process.

 

[RMerlin]

Hello everyone,

We just got the Asus AC68U for a small office.
12 people, 3x connect by Lan, 7x by 5Ghz Wireless and 2x by 2.4Ghz.
We already tested 3 firmware's:
- stock 45xx firmware
- stock 4887 (last FW from 27 April)
- Merlin Beta 3 (the last one from March)

With so many users, you will want to invest in a dedicated NAS rather than a USB disk shared by the router. Both for stability and performance reasons.

One thing I've noticed is that if you run a top or free command and if you keep your eye on the memory, this memory will decreased from the 256mb to 0mb during the copy.
So, if you copy 512Mb the RAM exhaust's.

Linux will use available memory for buffering. As soon a program actually needs that memory, the buffer will release it. That's normal behaviour. If you check with "free -m", you will see where the memory is being allocated.

 

[vapochilled]

Hi Merlin,

You're correct! It indeed cache everything.
I know that because I daily work with redhat clustering and Sun SPARC - Solaris.


My main question was if there is a relation between this. Some memory leak.... lets say....

 

[alkimio]

Hello everyone,

We just got the Asus AC68U for a small office.
12 people, 3x connect by Lan, 7x by 5Ghz Wireless and 2x by 2.4Ghz.
We already tested 3 firmware's:
- stock 45xx firmware
- stock 4887 (last FW from 27 April)
- Merlin Beta 3 (the last one from March)


With all of them we experience IO corruption issues when we try to copy 50mb of files.
System env:
- WD elements 1TB - USB3 port
- NTFS
- Using access with 5 users.



The fails are random. I've seen multiple problems like this across the web.
People that have IO issues during the copies.
Some people even changed the filesystem to EXT4 but the problem persisted.


Do you have a solution for this? Otherwise I will just send this router back while we're on time.
I've tried some samba tweaks over the /etc/smb.conf file but still no go.
Are these problems related to buffer?
Permissions?
One thing I've noticed is that if you run a top or free command and if you keep your eye on the memory, this memory will decreased from the 256mb to 0mb during the copy.
So, if you copy 512Mb the RAM exhaust's.


Any ideas guys? We're running out of time and the people in the office are getting tired because all the time we have the IO issue for a shirtty 20mb copy (in this case it is not even exhausting the memory ):
- the router webpage becomes inaccessible
- the shares don't work anymore
- im not able to reach the network printer


What can we do?
- tweak Samba?
- change to usb 2 port?
- use the NFS sharing instead provided with Merlin FW?
- change to Alpha 4 and try again?
- take this s*** back?

This same thing was happening to me.
My fix was installing merlin's alpha4 and enabling swap.
It looks like that memory is in fact an issue.
Please, try this and tell me!

 

[vapochilled]

BTW,

have you seen this in your logs?

Apr 4 12:41:12 dnsmasq-dhcp[688]: DHCPACK(br0) 192.168.2.20 08:60:6e:ca:61:53 JP-OGA
Apr 4 13:34:49 miniupnpd[780]: unsupported NAT-PMP version : 2
Apr 4 13:34:50 miniupnpd[780]: unsupported NAT-PMP version : 2
Apr 4 13:34:50 dnsmasq-dhcp[688]: DHCPREQUEST(br0) 192.168.2.105 7c:c3:a1:a1:75:7c
Apr 4 13:34:50 dnsmasq-dhcp[688]: DHCPACK(br0) 192.168.2.105 7c:c3:a1:a1:75:7c iMac-de-Bruno
Apr 4 13:52:04 dnsmasq-dhcp[688]: DHCPREQUEST(br0) 192.168.2.132 7c:c3:a1:a1:7c:92
Apr 4 13:52:04 dnsmasq-dhcp[688]: DHCPACK(br0) 192.168.2.132 7c:c3:a1:a1:7c:92 iMac-de-Tiago
Apr 4 13:52:39 miniupnpd[780]: unsupported NAT-PMP version : 2
Apr 4 13:52:40 dnsmasq-dhcp[688]: DHCPREQUEST(br0) 192.168.2.105 7c:c3:a1:a1:75:7c
Apr 4 13:52:40 dnsmasq-dhcp[688]: DHCPACK(br0) 192.168.2.105 7c:c3:a1:a1:75:7c iMac-de-Bruno
Apr 4 14:18:00 kernel: ------------[ cut here ]------------
Apr 4 14:18:00 kernel: WARNING: at fs/fs-writeback.c:1091 writeback_inodes_sb+0xc4/0xd4()
Apr 4 14:18:00 kernel: module: tun bf77c000 12723
Apr 4 14:18:00 kernel: module: nf_nat_sip bf775000 5586
Apr 4 14:18:00 kernel: module: nf_conntrack_sip bf769000 16679
Apr 4 14:18:00 kernel: module: nf_nat_h323 bf762000 5137
Apr 4 14:18:00 kernel: module: nf_conntrack_h323 bf751000 34844
Apr 4 14:18:00 kernel: module: nf_nat_rtsp bf74b000 3400
Apr 4 14:18:00 kernel: module: nf_conntrack_rtsp bf744000 4268
Apr 4 14:18:00 kernel: module: nf_nat_ftp bf73e000 1314
Apr 4 14:18:00 kernel: module: nf_conntrack_ftp bf737000 5131
Apr 4 14:18:00 kernel: module: ip6table_filter bf731000 893
Apr 4 14:18:00 kernel: module: ip6table_mangle bf72b000 1093
Apr 4 14:18:00 kernel: module: sr_mod bf722000 11507
Apr 4 14:18:00 kernel: module: cdrom bf713000 33318
Apr 4 14:18:00 kernel: module: zaurus bf70d000 2146
Apr 4 14:18:00 kernel: module: net1080 bf707000 2871
Apr 4 14:18:00 kernel: module: rndis_host bf700000 5193
Apr 4 14:18:00 kernel: module: cdc_ether bf6fa000 3224
Apr 4 14:18:00 kernel: module: asix bf6f1000 11629
Apr 4 14:18:00 kernel: module: usbnet bf6e7000 12057
Apr 4 14:18:00 kernel: module: mii bf6e1000 3484
Apr 4 14:18:00 kernel: module: usblp bf6d8000 9354
Apr 4 14:18:00 kernel: module: ohci_hcd bf6cd000 19288
Apr 4 14:18:00 kernel: module: ehci_hcd bf6be000 34220
Apr 4 14:18:00 kernel: module: xhci_hcd bf6a9000 53286
Apr 4 14:18:00 kernel: module: ufsd bf603000 595091
Apr 4 14:18:00 kernel: module: jnl bf5f6000 27927
Apr 4 14:18:00 kernel: module: ext2 bf5e0000 55581
Apr 4 14:18:00 kernel: module: ext4 bf597000 234233
Apr 4 14:18:00 kernel: module: jbd2 bf580000 52386
Apr 4 14:18:00 kernel: module: crc16 bf57a000 1081
Apr 4 14:18:00 kernel: module: ext3 bf553000 113117
Apr 4 14:18:00 kernel: module: jbd bf53e000 45524
Apr 4 14:18:00 kernel: module: mbcache bf536000 5156
Apr 4 14:18:00 kernel: module: usb_storage bf523000 34163
Apr 4 14:18:00 kernel: module: sg bf516000 21138
Apr 4 14:18:00 kernel: module: sd_mod bf50a000 23159
Apr 4 14:18:00 kernel: module: scsi_wait_scan bf504000 502
Apr 4 14:18:00 kernel: module: scsi_mod bf4d5000 114857
Apr 4 14:18:00 kernel: module: usbcore bf4aa000 108178
Apr 4 14:18:00 kernel: module: jffs2 bf489000 94871
Apr 4 14:18:00 kernel: module: zlib_deflate bf47f000 19990
Apr 4 14:18:00 kernel: module: nf_nat_pptp bf479000 1796
Apr 4 14:18:00 kernel: module: nf_conntrack_pptp bf473000 3739
Apr 4 14:18:00 kernel: module: nf_nat_proto_gre bf46d000 1047
Apr 4 14:18:00 kernel: module: nf_conntrack_proto_gre bf467000 3499
Apr 4 14:18:00 kernel: module: wl bf037000 3924613
Apr 4 14:18:00 kernel: module: igs bf02d000 12935
Apr 4 14:18:00 kernel: module: emf bf023000 16229
Apr 4 14:18:00 kernel: module: et bf00b000 63696
Apr 4 14:18:00 kernel: module: ctf bf000000 17519
Apr 4 14:18:00 kernel: Modules linked in: tun nf_nat_sip nf_conntrack_sip nf_nat_h323 nf_conntrack_h323 nf_nat_rtsp nf_conntrack_rtsp nf_nat_ftp nf_conntrack_ftp ip6table_filter ip6table_mangle sr_mod cdrom zaurus net1080 rndis_host cdc_ether asix usbnet mii usblp ohci_hcd ehci_hcd xhci_hcd ufsd(P) jnl ext2 ext4 jbd2 crc16 ext3 jbd mbcache usb_storage sg sd_mod scsi_wait_scan scsi_mod usbcore jffs2 zlib_deflate nf_nat_pptp nf_conntrack_pptp nf_nat_proto_gre nf_conntrack_proto_gre wl(P) igs(P) emf(P) et
Apr 4 14:18:00 kernel: [<c0043ff8>] (unwind_backtrace+0x0/0xf8) from [<c0061588>] (warn_slowpath_common+0x4c/0x64)
Apr 4 14:18:00 kernel: [<c0061588>] (warn_slowpath_common+0x4c/0x64) from [<c00615bc>] (warn_slowpath_null+0x1c/0x24)
Apr 4 14:18:00 kernel: [<c00615bc>] (warn_slowpath_null+0x1c/0x24) from [<c00e7358>] (writeback_inodes_sb+0xc4/0xd4)
Apr 4 14:18:00 kernel: [<c00e7358>] (writeback_inodes_sb+0xc4/0xd4) from [<bf603184>] (ufsd_get_sb+0x128/0x58c [ufsd])
Apr 4 14:18:00 kernel: [<bf603184>] (ufsd_get_sb+0x128/0x58c [ufsd]) from [<bf6033cc>] (ufsd_get_sb+0x370/0x58c [ufsd])
Apr 4 14:18:00 kernel: [<bf6033cc>] (ufsd_get_sb+0x370/0x58c [ufsd]) from [<bf60be6c>] (ufsd_readdir+0xbc/0x810 [ufsd])
Apr 4 14:18:00 kernel: [<bf60be6c>] (ufsd_readdir+0xbc/0x810 [ufsd]) from [<c00d8128>] (vfs_readdir+0x90/0xb0)
Apr 4 14:18:00 kernel: [<c00d8128>] (vfs_readdir+0x90/0xb0) from [<c00d82c4>] (sys_getdents64+0x6c/0xbc)
Apr 4 14:18:00 kernel: [<c00d82c4>] (sys_getdents64+0x6c/0xbc) from [<c003dac0>] (ret_fast_syscall+0x0/0x30)
Apr 4 14:18:00 kernel: ---[ end trace 895fb748bec642a2 ]---
Apr 4 14:52:39 miniupnpd[780]: unsupported NAT-PMP version : 2
Apr 4 14:56:05 dnsmasq-dhcp[688]: DHCPREQUEST(br0) 192.168.2.132 7c:c3:a1:a1:7c:92
Apr 4 14:56:05 dnsmasq-dhcp[688]: DHCPACK(br0) 192.168.2.132 7c:c3:a1:a1:7c:92 iMac-de-Tiago
Apr 4 14:57:45 kernel: ufsd:Incorrect dir/file of inode r=14585
Apr 4 14:57:57 dnsmasq-dhcp[688]: DHCPINFORM(br0) 192.168.2.28 08:60:6e:ca:67:9f
Apr 4 14:57:57 dnsmasq-dhcp[688]: DHCPACK(br0) 192.168.2.28 08:60:6e:ca:67:9f VanessaBarros
Apr 4 14:58:08 kernel: ufsd:Incorrect dir/file of inode r=145b2
Apr 4 14:59:25 kernel: ufsd:Incorrect dir/file of inode r=145a6
Apr 4 15:52:39 miniupnpd[780]: unsupported NAT-PMP version : 2
Apr 4 15:52:39 miniupnpd[780]: HTTP Connection from 192.168.2.105 closed unexpectedly
Apr 4 15:52:56 dnsmasq-dhcp[688]: DHCPREQUEST(br0) 192.168.2.132 7c:c3:a1:a1:7c:92
Apr 4 15:52:56 dnsmasq-dhcp[688]: DHCPACK(br0) 192.168.2.132 7c:c3:a1:a1:7c:92 iMac-de-Tiago
Apr 4 16:34:07 kernel: ufsd:Incorrect dir/file of inode r=14590
Apr 4 16:34:07 kernel: ufsd:Incorrect dir/file of inode r=145d6
Apr 4 16:52:39 miniupnpd[780]: unsupported NAT-PMP version : 2

 

[vapochilled]

How do you enable the swap?
With an additional pen drive?

 

[vapochilled]

any ideas?

 

[Adamm]

any ideas?

You need to create a swap partition when formatting the drive as ext. Many guides available and I'm almost positive you didn't use the search function.

 

[dipole]

You need to create a swap partition when formatting the drive as ext. Many guides available and I'm almost positive you didn't use the search function.

But unless I'm mistaken, you can't utilize swap on stock firmware correct? Only merlin? So if Asus didn't enable or require it for stock then why is it necessary? Just curious and new obviously....