BIG-IP i5820-DF reboot takes 15~ minutes
I have F5 VEs as well as physical F5s. My physicals are the i5820s. Whenever they are booting, it takes roughly 15-20 minutes to come up. With my VEs it takes maybe 5 minutes. After discussing with other F5 users as well as listening to an instructor, it seems that the boot length for my physicals is oddly long. Has anyone seen this?
Question to Chassis Temperature - Hi Limit (C)
We use snmp to monitor our f5 BIG-IP appliance (5250v). To configure the threshold we want to know if the "Hi Limit" is the right maximum value for the monitoring and the following temperature sensors/components: show /sys hardware: Chassis Temperature Status Index Lo Limit(C) Temp(degC) Hi Limit(C) Location 1 0 45 82 Main board HSBE transistor temp 2 0 28 62 Main board HSBE IC temperature 3 0 22 42 Main board inlet transistor tem 4 0 23 42 Main board inlet IC temperature 5 0 28 52 Main board outlet transistor te 6 0 30 52 Main board outlet IC temperatur 7 0 35 65 Power supply 1 meas. inlet tem 8 0 36 65 Power supply 2 meas. inlet tem 9 0 29 52 Main board near power supplies 10 0 33 82 Main board near B56821 temperat What happens when this value is reached? Is it just a information to know "this value is too high I should do anything" or would the f5 automatically turn off (for example)? I could not find any documentation, unfortunately. 😞
SNMP MIB: CPU Temperature
Hi guys, we saw, that the i-Series of the BigIP Appliances (i2000 and i4000) misses the CPU Temperature in the MIB Base. Also the tmsh: "show sys hardware" shows no information regarding that metric. This information used to be available in the former Hardware. Does somebody has a workaround? Or is it a feature? ;) Thanks, Rolf
Why am I seeing so many "Partial" SSL Hardware Accelerated transactions?
My ssl profile uses ciphers "default" show /ltm profile client-ssl myprofile gives this: (i've just recently reset the stats) SSL Hardware Acceleration Full 32.4K Partial 14.1K None (Software) 47 What are those Partial transactions? And how might I go about getting everything to be fully hw accelerated?
EUD not working on iSeries (i4000 & i5000)
Hello F5-ers, We always run the EUD test on new hardware to make sure there are no DOA issues. So projects don't frustrated if there is an issue and the engineer finds out when the device goes into production. Now we have four new iSeries devices, but we are unable to run the EUD test. Used these resources: https://support.f5.com/kb/en-us/products/big-ip_ltm/releasenotes/related/relnotes_eud.htmlrn_new_sf320 https://support.f5.com/kb/en-us/products/big-ip_ltm/releasenotes/related/eud-sf.html Ran the EUD first on a i4800, pick 'End User Diagnostics' from the grub boot loader. (connected via serial with putty) Preparing the Enter Diagnostic Software, Please Wait.... Setup Box Started alogd ... [OK] Started aserviced... [OK] Started arestd ... [OK] Loading QAT... ... ... Loading FPGA... HSB IO_12.0.11.1 loading successfully. To check for the latest version of the EUD, including details of function changes, go to: https://support.f5.com/kb/en-us/products/big-ip_ltm/releasenotes/related/relnotes_eud.html And then nothing happens anymore. On the display of the F5 itself it says "EUD in process, please wait" we let one F5's in this state for about 6 hours... nothing happened. Also ran the EUD from USB-stick, same result. There is only 1 EUD version available now (EUD_SF-v3.2.0) Then also tried to run EUD on a i5800 unit and i4600 units, all same result. Anyone else ran into this? Regards, Robert
i2800's L4 HTTP requests per second is lower than the 4000s
I would like to know why the L4 HTTP requests per second on the i2800 is lower than the 4000s? The rest of the specs on the iSeries is higher except for that item. Below are the specs from the hardware datasheet: BIG-IP i2800: L7 requests per second: 650K L4 connections per second: 250K L4 HTTP requests per second: 1M Maximum L4 concurrent connections: 14M Throughput: 10 Gbps L4/L7 BIG-IP 4000s: L7 requests per second: 425K L4 connections per second: 150K L4 HTTP requests per second: 1.25M Maximum L4 concurrent connections: 10M Throughput: 10 Gbps L4/L7
How to spec hardware for VE?
Dear all, this could an annoying question we may face selling Virtual Editions. I saw someone asked this question before but no one answered: https://devcentral.f5.com/questions/waf-throughput-sizing For example, if the apps team says: "I need 2000 SSL TPS, and expect 100,000 concurrent L4 connections, and I am running in a 1Gbps network environment because my switches have those interface ports." The question is, what hardware specs maintain 2000 SSL TPS and 100,000 concurrent L4 connections? In theory, CPU affects SSL performance while memory might affect concurrent connections since you need memory to maintain connection tables, etc. But in practice, what is the appropriate hardware spec? How would you have size it if you have specific performance metrics to meet? Looking at VE datasheet, we were given some range of performance metrics. (https://www.f5.com/pdf/products/big-ip-virtual-editions-datasheet.pdf) But it ranges for example, from 900 SSL RSA TPS (2K Keys) to 3,800 and the equipment tested for maximum performance is with Dell PowerEdge R620 with Intel Xeon CPU E5-2670 0 @ 2.6GHz and Intel 82599EB 10-Gigabit SFP+ NIC—configured for PCI pass-through with support for SR-IOV. How about using the hardware models as a gauge for VE since hardware models also state their hardware specs and related performance metrics? However, an old model of 5000 series runs Intel Xeon E3 Quad-Core 3.3Ghz chip by "tmsh show /sys hardware". I believe TMOS is very tightly-integrated with the hardware so perhaps a seemingly lower "Xeon" can run 5000 series?