Re: [Openvpn-devel] Slight change to buildbot t_client.sh.in & t_client.rc
On 06/10/16 18:01, Gert Doering wrote: > Hi, > > On Thu, Oct 06, 2016 at 04:47:21PM +0100, debbie10t wrote: >> As I am on possibly the cheapest internet in the UK and "-i 1000" >> solves the problem, i expect simply including say "-i 2000" in >> t_client.sh.in will be sufficient for just about anywhere, unless >> you try to run a buildslave in the jungles of Borneo ! > > To repeat this: *no*. We're not including -i 2000 by default in the > script, because that will slow down things for all testers (and t_client > takes enough time as it is, I run it interactively often enough). > > gert > Sorry to drag this up again ... I have done a little more testing and simply adding -i 100 allows fping to succeed on my buildbots. (more or less, some pings fail but target is alive) Also note; over 60 pings (my tests) this adds about one second, so for 20 pings (t_client) this will add ~300ms per fping test. These timings are very approximate because using -c 20 -p 250 -i 50 ran faster than -c 20 -p 250 -i 25 by ~300ms, see last tests attached. Regards == TESTS: Running three bots *concurrently* (arch-64, cos7-64, gtoo-64): -i 25 (default): [root@arch-hyv-live-64 t_client-arch-hyv-live-64-20161020-005508]# fping -b 3000 -s -r 1 -q -C 60 -p 250 -i 25 10.194.6.1 10.194.0.1 10.194.6.1 : 204.64 205.65 205.11 204.47 204.49 205.04 204.96 204.23 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 10.194.0.1 : - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 2 targets 1 alive 1 unreachable 0 unknown addresses 2 timeouts (waiting for response) 120 ICMP Echos sent 8 ICMP Echo Replies received 0 other ICMP received 204 ms (min round trip time) 204 ms (avg round trip time) 205 ms (max round trip time) 15.324 sec (elapsed real time) [root@cos7-hyv-live-64 t_client-cos7-hyv-live-64-20161019-222647]# fping -b 3000 -s -q -C 60 -p 250 -i 25 10.194.6.1 10.194.0.1 10.194.6.1 : - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 214.32 215.88 215.79 10.194.0.1 : - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 2 targets 1 alive 1 unreachable 0 unknown addresses 2 timeouts (waiting for response) 120 ICMP Echos sent 3 ICMP Echo Replies received 0 other ICMP received 214 ms (min round trip time) 215 ms (avg round trip time) 215 ms (max round trip time) 15.327 sec (elapsed real time) gtoo-hyv-live-64 t_client-gtoo-hyv-live-64-20161019-222726 # fping -b 3000 -s -q -C 60 -p 250 -i 25 10.194.6.1 10.194.0.1 10.194.6.1 : 205.06 205.26 205.79 205.57 205.03 205.29 204.92 204.79 204.51 204.89 204.12 204.78 204.88 204.79 204.58 205.15 205.10 204.93 204.77 204.66 205.27 204.27 204.85 204.66 205.43 204.59 205.63 204.46 205.32 204.04 205.40 204.27 204.62 205.00 204.90 205.14 206.55 207.66 204.72 204.46 204.78 204.69 204.29 204.57 205.09 204.39 204.53 205.44 205.19 204.67 205.22 205.45 205.07 204.23 204.93 204.82 205.74 204.43 205.31 204.69 10.194.0.1 : - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 2 targets 1 alive 1 unreachable 0 unknown addresses 1 timeouts (waiting for response) 120 ICMP Echos sent 60 ICMP Echo Replies received 0 other ICMP received 204 ms (min round trip time) 204 ms (avg round trip time) 207 ms (max round trip time) 15.316 sec (elapsed real time) * Verses * Running three bots *concurrently* (arch-64, cos7-64, gtoo-64): -i 100 [root@arch-hyv-live-64 t_client-arch-hyv-live-64-20161020-005508]# fping -b 3000 -s -r 1 -q -C 60 -p 250 -i 100 10.194.6.1 10.194.0.1 10.194.6.1 : 204.55 206.03 205.79 204.62 205.60 222.08 221.34 221.46 222.40 222.19 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 10.194.0.1 : 205.62 204.56 204.32 204.40 204.01 205.02 204.73 205.06 205.11 204.53 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 2 targets 2 alive 0 unreachable 0 unknown addresses 2 timeouts (waiting for response) 120 ICMP Echos sent 20 ICMP Echo Replies received 0 other ICMP received 204 ms (min round trip time) 209 ms (avg round trip time) 222 ms (max round trip time) 15.471 sec (elapsed real time) [root@cos7-hyv-live-64 t_client-cos7-hyv-live-64-20161019-222647]# fping -b 3000 -s -q -C 60 -p 250 -i 100 10.194.6.1 10.194.0.1 10.194.6.1 : - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Re: [Openvpn-devel] Slight change to buildbot t_client.sh.in & t_client.rc
Hi, On Thu, Oct 06, 2016 at 04:47:21PM +0100, debbie10t wrote: > As I am on possibly the cheapest internet in the UK and "-i 1000" > solves the problem, i expect simply including say "-i 2000" in > t_client.sh.in will be sufficient for just about anywhere, unless > you try to run a buildslave in the jungles of Borneo ! To repeat this: *no*. We're not including -i 2000 by default in the script, because that will slow down things for all testers (and t_client takes enough time as it is, I run it interactively often enough). gert -- USENET is *not* the non-clickable part of WWW! //www.muc.de/~gert/ Gert Doering - Munich, Germany g...@greenie.muc.de fax: +49-89-35655025g...@net.informatik.tu-muenchen.de signature.asc Description: PGP signature -- Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel
Re: [Openvpn-devel] Slight change to buildbot t_client.sh.in & t_client.rc
On 06/10/16 14:10, debbie10t wrote: > > > On 06/10/16 13:25, Arne Schwabe wrote: >> >> >> Am 06.10.16 um 14:17 schrieb Gert Doering: >>> Hi, >>> >>> On Thu, Oct 06, 2016 at 01:02:57PM +0100, debbie10t wrote: -R Instead of using all-zeros as the packet data, generate random bytes. Use to defeat, e.g., link data compression >>> -R is unlikely to have an effect here, as test 6 is run without >>> compression anyway (to the contrary, it would *fail* your tests that >>> succeed today due to compression :-) ) >>> >>> >> You just giving a good point to add -R anyway! >> >> Arne >> > > Turns out -R is not currently supported by CentOS7: > $ fping -v > fping: Version 3.10 > fping: comments to da...@schweikert.ch > > $ fping -R -b 3000 -c 10 -p 250 10.194.6.1 10.194.0.1 > fping: invalid option -- 'R' > > Arch-linux: > $ fping -v > fping: Version 3.13 > > So I guess it is not worth over complicating t_client for that. > > > === > > However, -i appears to be commonly supported and this is the effect for me: > > --- > > [root@arch-hyv-live-64 t_client-arch-hyv-live-64-20161004-210205]# fping > -b 3000 -c 10 -p 250 10.194.6.1 10.194.0.1 > 10.194.6.1 : [0], 3028 bytes, 551 ms (551 avg, 66% loss) > > 10.194.6.1 : xmt/rcv/%loss = 10/1/90%, min/avg/max = 551/551/551 > 10.194.0.1 : xmt/rcv/%loss = 10/0/100% > > --- > > [root@arch-hyv-live-64 t_client-arch-hyv-live-64-20161004-210205]# fping > -i 1000 -b 3000 -c 10 -p 250 10.194.6.1 10.194.0.1 > 10.194.6.1 : [0], 3028 bytes, 550 ms (550 avg, 0% loss) > 10.194.0.1 : [0], 3028 bytes, 552 ms (552 avg, 0% loss) > 10.194.6.1 : [1], 3028 bytes, 550 ms (550 avg, 0% loss) > 10.194.0.1 : [1], 3028 bytes, 552 ms (552 avg, 0% loss) > 10.194.6.1 : [2], 3028 bytes, 549 ms (550 avg, 0% loss) > 10.194.0.1 : [2], 3028 bytes, 552 ms (552 avg, 0% loss) > 10.194.6.1 : [3], 3028 bytes, 549 ms (550 avg, 0% loss) > 10.194.0.1 : [3], 3028 bytes, 551 ms (552 avg, 0% loss) > 10.194.6.1 : [4], 3028 bytes, 553 ms (550 avg, 0% loss) > 10.194.0.1 : [4], 3028 bytes, 573 ms (556 avg, 0% loss) > 10.194.6.1 : [5], 3028 bytes, 552 ms (551 avg, 0% loss) > 10.194.0.1 : [5], 3028 bytes, 550 ms (555 avg, 0% loss) > 10.194.6.1 : [6], 3028 bytes, 552 ms (551 avg, 0% loss) > 10.194.0.1 : [6], 3028 bytes, 558 ms (555 avg, 0% loss) > 10.194.6.1 : [7], 3028 bytes, 551 ms (551 avg, 0% loss) > 10.194.0.1 : [7], 3028 bytes, 548 ms (554 avg, 0% loss) > 10.194.6.1 : [8], 3028 bytes, 555 ms (551 avg, 0% loss) > 10.194.0.1 : [8], 3028 bytes, 557 ms (555 avg, 0% loss) > 10.194.6.1 : [9], 3028 bytes, 563 ms (553 avg, 0% loss) > 10.194.0.1 : [9], 3028 bytes, 570 ms (556 avg, 0% loss) > > 10.194.6.1 : xmt/rcv/%loss = 10/10/0%, min/avg/max = 549/553/563 > 10.194.0.1 : xmt/rcv/%loss = 10/10/0%, min/avg/max = 548/556/573 > > --- > > So, if you want the fping's to work adding configurable "-i $interval" > to t_client.sh.in/t_client.rc looks like a simple enough solution. > > Thanks A little more thought on this: As I am on possibly the cheapest internet in the UK and "-i 1000" solves the problem, i expect simply including say "-i 2000" in t_client.sh.in will be sufficient for just about anywhere, unless you try to run a buildslave in the jungles of Borneo ! proost -- Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot ___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel
Re: [Openvpn-devel] Slight change to buildbot t_client.sh.in & t_client.rc
Am 06.10.16 um 14:17 schrieb Gert Doering: > Hi, > > On Thu, Oct 06, 2016 at 01:02:57PM +0100, debbie10t wrote: >> -R >> Instead of using all-zeros as the packet data, generate random bytes. >> Use to defeat, e.g., link data compression > -R is unlikely to have an effect here, as test 6 is run without > compression anyway (to the contrary, it would *fail* your tests that > succeed today due to compression :-) ) > > You just giving a good point to add -R anyway! Arne -- Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot ___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel
Re: [Openvpn-devel] Slight change to buildbot t_client.sh.in & t_client.rc
Hi, On Thu, Oct 06, 2016 at 01:02:57PM +0100, debbie10t wrote: > -R > Instead of using all-zeros as the packet data, generate random bytes. > Use to defeat, e.g., link data compression -R is unlikely to have an effect here, as test 6 is run without compression anyway (to the contrary, it would *fail* your tests that succeed today due to compression :-) ) gert -- USENET is *not* the non-clickable part of WWW! //www.muc.de/~gert/ Gert Doering - Munich, Germany g...@greenie.muc.de fax: +49-89-35655025g...@net.informatik.tu-muenchen.de signature.asc Description: PGP signature -- Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel
[Openvpn-devel] Slight change to buildbot t_client.sh.in & t_client.rc
Hi My buildbots always to fail "fping 4 & 6 -b 3000" tests probably due to my ISP. I have found that simply adding "-R -i 5000" to fping 4 & 6 solves my problem completely. -R Instead of using all-zeros as the packet data, generate random bytes. Use to defeat, e.g., link data compression -i n The minimum amount of time (in milliseconds) between sending a ping packet to any target (default is 25) As gert does not want to increase the time taken for bots which do not have this problem, I would like to propose that "-R -i $interval" is added to t_client.sh.in and $interval be sourced from t_client.rc, where it can be configured for each slave uniquely. (default=25) I am still wrestling with git but I could try to make this change myself, if it gets a thumbs up. As it is such a low priority (effects only me) it would seem like a suitable candidate for me to try. Otherwise, if anybody prefers to apply this for me, I would be most grateful :-) (if it is acceptable) Thanks -- Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot ___ Openvpn-devel mailing list Openvpn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openvpn-devel