Welp, still too load for me. At 30cm the noise is 41db. My whole 32U rack is 31db with 3 servers (water cooling with 6x low RPM noctua + a few EXOS&WD GOLD HDD) and 1980 (with noctua)
Last edited:
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
Now I have modified it with Noctua fans, but I didn't liked that that temp of front sensors were around 60 degree, so I added two small (30x30x15) fans right near the big heatsink with Noctua noise reduction, now it's not louder, but cooler.
Also
If you found this post and you planning to modify DX7050SX-s with 495W powersupply
They using different colors for PWM and Sense in PSU
The correct color for PSU is
- Ground: Arista Black = Noctua Black
- +12V: Arista Red = Noctua Yellow
- RPM: Arista Light Blue = Noctua Green
- PWM: Arista Yellow = Noctua Blue
Can some clarify what using third party transceivers means? Does mean the same type of transceiver, but from Mellanox or Intel instead of Arista. Or can I use a type not listed in the Arista PDF for the switch? For example I already have a bunch of SPTSBP2CLCKV INTEL 100G-CWDM4 transceivers I was going to use with a Celestica, but I changed my mind want something off the shelf so to speak. I know the 100G-CWDM4 negotiates at 40Gbe because I currently have a few computers directly linked together, but I'm unsure if the Arista 7050QX will recognize them. I'm also tempted by the Arista Networks DCS-7050SX-72Q , but I'm unsure which CPU is in those and I'd like to avoid the Intel C2xxx with the bug.
As far as the 7050QX that's going to come down to whether or not you can stick a QSFP28 module into a QSFP port and have it work. I don't know the rules on mixing QSFP and QSFP28.
As far as the CPU, my 7050QX-32S has an AMD chip, the GX-420CA SOC. I can't readily find what CPU is in the 7050SX series.
I currently have the QSFP28 in QSFP Mellanox NICs so I'm confident that is ok generally speaking, but CWDM4 isn't listed for the 7050QX so that's where my confusion arises. I'm not sure if a switch knows or cares what tech the transceiver is using. I'm guessing the list in the PDF is just what Arista offers and has tested with the switch.
I tried searching what CPU is in the 7050SX series the PDF manual I found states x86 quad core if I remember correctly, I closed the PDF.
Thanks for the quick response.
Over in the Power Consumption thread, looks like @john389 has a 7050SX2-72Q, which may have a different board than the 7050SX-72Q, maybe he'd be willing to drop into bash on his switch and pull the info from /proc/cpuinfo.
Arista DCS-7050SX2-72Q:
The SX2 has the Trident2+ ASIC while the SX has the Trident2.
If I remember correctly there were some differences between the two. Trident2+ is the one with RIOT support, see VXLAN Routing | Cumulus Linux 4.0 . I think there was something else as well, but I can't recall at the moment.
Code:
#bash cat /proc/cpuinfo
processor : 0
vendor_id : AuthenticAMD
cpu family : 22
model : 48
model name : AMD GX-424CC SOC with Radeon(TM) R5E Graphics
stepping : 1
microcode : 0x7030109
cpu MHz : 2402.905
cache size : 2048 KB
physical id : 0
siblings : 4
core id : 0
cpu cores : 4
apicid : 0
initial apicid : 0
fpu : yes
fpu_exception : yes
cpuid level : 13
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good acc_power nopl nonstop_tsc cpuid extd_apicid aperfmperf pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt topoext perfctr_nb bpext ptsc perfctr_llc hw_pstate ssbd ibpb vmmcall bmi1 xsaveopt arat npt lbrv svm_lock nrip_save tsc_scale flushbyasid decodeassists pausefilter pfthreshold overflow_recov
bugs : fxsave_leak sysret_ss_attrs null_seg spectre_v1 spectre_v2 spec_store_bypass
bogomips : 4799.86
TLB size : 1024 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts ttp tm 100mhzsteps hwpstate acc_power [13]
processor : 1
vendor_id : AuthenticAMD
cpu family : 22
model : 48
model name : AMD GX-424CC SOC with Radeon(TM) R5E Graphics
stepping : 1
microcode : 0x7030109
cpu MHz : 2402.169
cache size : 2048 KB
physical id : 0
siblings : 4
core id : 1
cpu cores : 4
apicid : 1
initial apicid : 1
fpu : yes
fpu_exception : yes
cpuid level : 13
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good acc_power nopl nonstop_tsc cpuid extd_apicid aperfmperf pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt topoext perfctr_nb bpext ptsc perfctr_llc hw_pstate ssbd ibpb vmmcall bmi1 xsaveopt arat npt lbrv svm_lock nrip_save tsc_scale flushbyasid decodeassists pausefilter pfthreshold overflow_recov
bugs : fxsave_leak sysret_ss_attrs null_seg spectre_v1 spectre_v2 spec_store_bypass
bogomips : 4799.86
TLB size : 1024 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts ttp tm 100mhzsteps hwpstate acc_power [13]
processor : 2
vendor_id : AuthenticAMD
cpu family : 22
model : 48
model name : AMD GX-424CC SOC with Radeon(TM) R5E Graphics
stepping : 1
microcode : 0x7030109
cpu MHz : 2400.662
cache size : 2048 KB
physical id : 0
siblings : 4
core id : 2
cpu cores : 4
apicid : 2
initial apicid : 2
fpu : yes
fpu_exception : yes
cpuid level : 13
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good acc_power nopl nonstop_tsc cpuid extd_apicid aperfmperf pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt topoext perfctr_nb bpext ptsc perfctr_llc hw_pstate ssbd ibpb vmmcall bmi1 xsaveopt arat npt lbrv svm_lock nrip_save tsc_scale flushbyasid decodeassists pausefilter pfthreshold overflow_recov
bugs : fxsave_leak sysret_ss_attrs null_seg spectre_v1 spectre_v2 spec_store_bypass
bogomips : 4799.86
TLB size : 1024 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts ttp tm 100mhzsteps hwpstate acc_power [13]
processor : 3
vendor_id : AuthenticAMD
cpu family : 22
model : 48
model name : AMD GX-424CC SOC with Radeon(TM) R5E Graphics
stepping : 1
microcode : 0x7030109
cpu MHz : 2402.485
cache size : 2048 KB
physical id : 0
siblings : 4
core id : 3
cpu cores : 4
apicid : 3
initial apicid : 3
fpu : yes
fpu_exception : yes
cpuid level : 13
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good acc_power nopl nonstop_tsc cpuid extd_apicid aperfmperf pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt topoext perfctr_nb bpext ptsc perfctr_llc hw_pstate ssbd ibpb vmmcall bmi1 xsaveopt arat npt lbrv svm_lock nrip_save tsc_scale flushbyasid decodeassists pausefilter pfthreshold overflow_recov
bugs : fxsave_leak sysret_ss_attrs null_seg spectre_v1 spectre_v2 spec_store_bypass
bogomips : 4799.86
TLB size : 1024 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts ttp tm 100mhzsteps hwpstate acc_power [13]If I remember correctly there were some differences between the two. Trident2+ is the one with RIOT support, see VXLAN Routing | Cumulus Linux 4.0 . I think there was something else as well, but I can't recall at the moment.
Last edited:
Thanks, I'm gonna look at those switches too, having 6 QSFP+ ports is enough for me for now I think and having SFP+ ports would be nice too. This seems like a more versatile option than just QSFP+.
As I understand the way these are built, you have one board that is CPU and storage for management, then a separate board that’s attached by PCIe or other link where the ASIC lives and the management board just sends commands to that. So not uncommon to see the same generation CPU board driving a newer ASIC board.
Anyone know if setting the fan override percentage convinces the switch to ignore the "fan speed too low" warnings and subsequent auto-shutdown?
I've put Noctuas in, but it started complaining about low fan speeds. Assuming this is because they're reporting back their actual RPM which doesn't align with what the switch is expecting for the PWM.
I set the speed override to 30%, and it's been running stable ever since with no complaints, but if I'm able to bump it to 50-60% without worry of the switch shutting itself down that'd be great as it's still plenty quiet.
I've put Noctuas in, but it started complaining about low fan speeds. Assuming this is because they're reporting back their actual RPM which doesn't align with what the switch is expecting for the PWM.
I set the speed override to 30%, and it's been running stable ever since with no complaints, but if I'm able to bump it to 50-60% without worry of the switch shutting itself down that'd be great as it's still plenty quiet.