Greetings!!
We are trying to install LEDE/OpenWRT in Linksys WRT1900ACSv2. We found the installation URL from the hardware list (https://openwrt.org/toh/views/toh_fwdownload?dataflt[Model*~]=wrt1900acs&dataflt[Brand*~]=linksys). However the page only shows one option for 1900ACS while for 1900AC it shows separate links for v1 and v2.
Is there only one common image for both version 1 and version 2 of WRT1900ACS ?
Thank You
They both (v1 or v2), for acs, use the same image.
Thank You !
Thank you!
Hello David, Greetings.
I work with Abin, who had asked the earlier question. Thanks. We flashed the Linksys rotuer with the firmware. The basic functionality i.e., enabling bring up with OVS, and enabling switching using OpenvSwitch works fine. However we have challenges.
Our (abin & myself) goal is to use OVS support over the LAN ports of Linksys Wrt 1900ACS v2, and create a high speed OVS enabled switch. We observe high latency when OVS is enabled. i.e., a ping between two end systems with the OVS enabled results in a round trip delay of about 400+micro seconds (0.4+ ms). Without OVS, the base switching silicon gives a round trip delay of about 180+micro seconds (0.18+ms). Our expectation was OVS to be give as close to basic switching silicon speed. Do you have any pointer/suggestion which can reduce the latency. or in other words how to increse ovs kernel forwarding speed?
Also we observed the ring buffers rx 128 and tx as 532. Any pointers as how to increase the rx buffers? we thought Linksys 1900ACS being a powerful device will have large ring buffers, but got plesantly surprised.
thank you
Manikantan
Hello, Greetings.
I work with Abin, who had asked the earlier question. Thanks. We flashed the Linksys rotuer with the firmware. The basic functionality i.e., enabling bring up with OVS, and enabling switching using OpenvSwitch works fine. However we have challenges.
Our (abin & myself) goal is to use OVS support over the LAN ports of Linksys Wrt 1900ACS v2, and create a high speed OVS enabled switch. We observe high latency when OVS is enabled. i.e., a ping between two end systems with the OVS enabled results in a round trip delay of about 400+micro seconds (0.4+ ms). Without OVS, the base switching silicon gives a round trip delay of about 180+micro seconds (0.18+ms). Our expectation was OVS to be give as close to basic switching silicon speed. Do you have any pointer/suggestion which can reduce the latency. or in other words how to increse ovs kernel forwarding speed?
Also we observed the ring buffers rx 128 and tx as 532. Any pointers as how to increase the rx buffers? we thought Linksys 1900ACS being a powerful device will have large ring buffers, but got plesantly surprised.
thank you
Manikantan
You don't have to post the same replies to multiple people, just mention the user(s) via @< username > (i.e. @manikantan @abinmathew2) and the user will be notified they've been mentioned in a post =]
I have no experience with OVS, but 400ms does seem quite excessive. If you haven't already done so, I would utilize forum search until someone more knowledgeable is able to reply.
That's not 400ms, that 0.4ms which is less than 1ms.
thank you for the pointers.. will dig further the forum post, and wait for reply.
As @davidc502 metioned, the round trip time is now 0.4 ms, which ideally must be less than or equal to 0.2ms or less. TP-Link 1043ND gives 0.2ms. We need to figure out if there is a way to reduce it.
Also our application generates burst of data. May be a large tx/rx ring buffer size is required. We have to figure out the possibility for increasing this.
Thanks for your support. Thank you @davidc502
Sorry, I can't be of more help as this is in an area I'm unfamiliar with.
The old forum was horrible with searches, and with a shell of the forum being stood up in its place now, I'm not going to find what I'm looking for. However, I have heard of people attempting to tweak buffer sizes, but have no idea how it was done.
EDIT
The first place I looked was sysctl.conf. However, I don't see the usual options to adjust wmem max.
Thank you @davidc502 for your time and replies. Much appreciated.
Will continue looking out for a suitable solution.
Wishing you a good weekend.
Regards
Manikantan
Out of curiosity, why does the extra 200μs matter?
Thank you for asking the question @JW0914
Now the measured .4 ms round-trip (ping) latency indicates (assuming symmetric network behavior) a one-way latency of .2 ms. This is with Host 1<--> Linksys WRT1900 ACS <---> Host 2. A back to back Host1<---> Host2 round-trip ping latency is about 0.1ms, one way will be .05ms. This leads to conclude the latency overhead added by one hop (Linksys WRT1900 ACS ) is 0.15 ms.
In my setup, i am looking forward to a network with 4 hops, such as Host1 <-> S1<->S2<->S3<->S4<->H2. Now 4 switches in between each adding .15 ms will add to a total of .6ms latency.
My application (a mobile LTE network testbed) expects data transfers and response back within certain time limits such as 3 ms. Hence as much reduced latency will help.
I do agree, slightly higher latency might not be a serious deterrent for my setup, but packet drops cause issues. If forwarding is faster, less latency, imply less buffer overrun, and thus no packet drops.
We burst 14 packets every 1 ms, 14 packets (89,600 bits) get generated within a short span say within 15 microseconds, the rate is about 500 Mbps (for this 15-microsecond duration). Buffer size to handle the burst is critical for me. Hence looking for pointers that could help to increase on TX/RX ring buffers/buffer sizes, i.e. there is no loss in a network having 4 hops.
thank you
Manikantan
@abinmathew2
Did anyone try to connect a storage device with sata port multiplier to this router? Is it working?
I'm trying to connect an Areca ARC-5040 DAS and I get kernel errors:
<6>[ 154.685985] ata2: SATA link down (SStatus 0 SControl 300)
<6>[ 156.135976] ata2: SATA link down (SStatus 100 SControl 300)
<3>[ 158.535973] ata2: COMRESET failed (errno=-32)
<4>[ 158.540356] ata2: reset failed (errno=-32), retrying in 8 secs
<6>[ 167.355986] ata2: SATA link down (SStatus 101 SControl 300)
<6>[ 168.885983] ata2: SATA link down (SStatus 100 SControl 300)
<3>[ 171.345979] ata2: COMRESET failed (errno=-32)
<4>[ 171.350359] ata2: reset failed (errno=-32), retrying in 8 secs
<6>[ 179.815978] ata2: SATA link down (SStatus 0 SControl 300)
<6>[ 180.675985] ata2: SATA link down (SStatus 100 SControl 300)
<6>[ 181.695978] ata2: SATA link down (SStatus 100 SControl 300)
<6>[ 182.505984] ata2: SATA link down (SStatus 100 SControl 300)
<6>[ 183.775979] ata2: SATA link down (SStatus 100 SControl 300)
<6>[ 184.545984] ata2: SATA link down (SStatus 100 SControl 300)
<6>[ 185.416028] ata2: SATA link down (SStatus 100 SControl 300)
<3>[ 187.845976] ata2: COMRESET failed (errno=-32)
<4>[ 187.850364] ata2: reset failed (errno=-32), retrying in 8 secs
<6>[ 197.155980] ata2: SATA link down (SStatus 100 SControl 300)
<3>[ 199.425975] ata2: COMRESET failed (errno=-32)
<4>[ 199.430361] ata2: reset failed (errno=-32), retrying in 8 secs
[...]
<3>[ 770.865973] ata2: COMRESET failed (errno=-32)
<4>[ 770.870360] ata2: reset failed (errno=-32), retrying in 8 secs
<6>[ 779.515985] ata2: SATA link down (SStatus 101 SControl 300)
<6>[ 780.275979] ata2: SATA link down (SStatus 100 SControl 300)
<3>[ 780.281595] ata2: EH pending after 5 tries, giving up
The same errors appear with OpenWrt official snapshot image.
The same DAS with the same eSATA cable works fine with WRT-1900AC v1.
<6>[130900.453758] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl F300)
<6>[130900.461286] ata1.15: Port Multiplier 1.2, 0x197b:0x0325 r193, 8 ports, feat 0xf/0x1f
<6>[130903.119327] ata1.00: ATA-6: WDC WD50ARC-5040-VOL#01, 0100 AX, max UDMA/133
<6>[130903.126400] ata1.00: 976773168 sectors, multi 0: LBA48
<6>[130903.132140] ata1.00: configured for UDMA/133
<6>[130903.139272] ata1.01: ATA-6: Areca Archive, 0100 AX, max UDMA/133
<6>[130903.145606] ata1.01: 23437498368 sectors, multi 0: LBA48
<6>[130903.151613] ata1.01: configured for UDMA/133
<5>[130903.157787] scsi 0:0:0:0: Direct-Access ATA WDC WD50ARC-5040 n/a PQ: 0 ANSI: 5
<5>[130903.168054] sd 0:0:0:0: [sda] 976773168 512-byte logical blocks: (500 GB/466 GiB)
<5>[130903.177019] sd 0:0:0:0: [sda] Write Protect is off
<7>[130903.181941] sd 0:0:0:0: [sda] Mode Sense: 00 3a 00 00
<5>[130903.182126] scsi 0:1:0:0: Direct-Access ATA Areca Archive n/a PQ: 0 ANSI: 5
<5>[130903.183965] sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
<5>[130903.200608] sd 0:1:0:0: [sdb] 23437498368 512-byte logical blocks: (12.0 TB/10.9 TiB)
<5>[130903.209265] sd 0:1:0:0: [sdb] Write Protect is off
<7>[130903.214185] sd 0:1:0:0: [sdb] Mode Sense: 00 3a 00 00
<5>[130903.214840] sd 0:1:0:0: [sdb] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
<5>[130903.686769] sd 0:1:0:0: [sdb] Attached SCSI disk
<5>[130903.688694] sd 0:0:0:0: [sda] Attached SCSI disk
<6>[130966.139077] BTRFS info (device sda): flagging fs with big metadata feature
<6>[130966.146110] BTRFS info (device sda): using free space tree
<6>[130966.151701] BTRFS info (device sda): has skinny extents
<6>[130974.284195] BTRFS info (device sdb): flagging fs with big metadata feature
<6>[130974.291227] BTRFS info (device sdb): using free space tree
<6>[130974.296852] BTRFS info (device sdb): has skinny extents