[solved] Usually showing "cannot display the web page"

redstrike

Well-Known Member
#21
if so, you can enable AIO instead of sendfile to see if any improvement.

Server->Tuning->Use Asynchronized disk I/O (AIO)
Thanks NiteWave, i will try AIO when i have owned a license. Maybe, it wouldn't help because the limit of VPS itself. So i am going to add more node.
 

redstrike

Well-Known Member
#22
if so, you can enable AIO instead of sendfile to see if any improvement.

Server->Tuning->Use Asynchronized disk I/O (AIO)
You mean when we enable AIO, we must disable "Use sendfile()" to allow AIO starting to work? I have tried 2-CPU trial license, but that time i didn't disable sendfiles option and the result wasn't better.

After that trial time, i understood that i got big disk I/O problem when other sites was using my media url to stream on their site. And come to now, i feel that it is difficult to access realtime stats page while peak hours so that is very annoy, viewing server logs is difficult too.

LiteSpeed is a great web server, i hope via my case. LiteSpeed Team can figure something out and optimize it far better.
 

redstrike

Well-Known Member
#24
sendfile should be enabled along with AIO.
Maybe you are using PHP script to streaming files, it is low efficient, if so, you should consider using internal redirect feature.
http://www.litespeedtech.com/support/wiki/doku.php?id=litespeed:wiki:feature:internal_redirect
I am not sure i am using PHP script to streaming files or not. My scripts create a playlist file (asx format) with full URL to media file. Can you check for it me? http://vn2rap.com/m/#Play,20200 --- http://vn2rap.com/m/asx.php?type=1&id=20200

I am using PHP. How can i use internal redirect?
 

redstrike

Well-Known Member
#26
the 2 URLs not accessible or empty page
This is the main thing: http://vn2rap.com/m/#Play,20200
Player streaming asx create by this link: http://vn2rap.com/m/asx.php?type=1&id=20200 it doesn't show directly on browser, you should use some downloader to download. It will show it's content.

Does it shows that: "cannot display ..." ? You should try to refresh after a coupe seconds. This issues is what i have mentioned. In cases, do you need to access SSH, and litespeed cp to messure my problem?
 

NiteWave

Administrator
#29
received.
it's a OpenVZ container.
Code:
~>date;cat /proc/user_beancounters;date
Tue Sep  6 13:06:46 ICT 2011
Version: 2.5
       uid  resource                     held              maxheld              barrier                limit              failcnt
   244074:  kmemsize                 11820364             28828390           2147483646           2147483646                    0
            lockedpages                     0                    8                 1000                 1000                    5
            privvmpages                 97081               166263               196608               196608                    0
            shmpages                    11989                52035                98304                98304                    0
            dummy                           0                    0                    0                    0                    0
            numproc                       111                  304                  500                  500                    0
            physpages                   26472                75044                    0           2147483647                    0
            vmguarpages                     0                    0               196608               196608                    0
            oomguarpages                26810                75045               196608               196608                    0
            numtcpsock                     45                 1000                 1000                 1000                 7035
            numflock                       21                  137                 1000                 1000                    0
            numpty                          1                    3                  500                  500                    0
            numsiginfo                     53                  254                 1000                 1000                    0
            tcpsndbuf                 3144432             10843328              3145728              4194304          79847400259
            tcprcvbuf                  261776              5311712              2097152              3145728                    0
            othersockbuf                94376              4708440              2097152              3145728                 5084
            dgramrcvbuf                     0                16944             80530432            113298432                    0
            numothersock                   55                  626                 5000                 5000                    0
            dcachesize                4688185              7027826              9216000              9216000                    0
            numfile                      2142                 4118                24000                24000                    0
            dummy                           0                    0                    0                    0                    0
            dummy                           0                    0                    0                    0                    0
            dummy                           0                    0                    0                    0                    0
            numiptent                      38                  200                 1500                 1500                    0
Tue Sep  6 13:06:46 ICT 2011
~>
looking at the failcnt column, there are a lot of failures.
please increase barrier/limit, especially tcpsndnif at host first.
let's see if any improvement.
 

redstrike

Well-Known Member
#30
received.
it's a OpenVZ container.
Code:
~>date;cat /proc/user_beancounters;date
Tue Sep  6 13:06:46 ICT 2011
Version: 2.5
       uid  resource                     held              maxheld              barrier                limit              failcnt
   244074:  kmemsize                 11820364             28828390           2147483646           2147483646                    0
            lockedpages                     0                    8                 1000                 1000                    5
            privvmpages                 97081               166263               196608               196608                    0
            shmpages                    11989                52035                98304                98304                    0
            dummy                           0                    0                    0                    0                    0
            numproc                       111                  304                  500                  500                    0
            physpages                   26472                75044                    0           2147483647                    0
            vmguarpages                     0                    0               196608               196608                    0
            oomguarpages                26810                75045               196608               196608                    0
            numtcpsock                     45                 1000                 1000                 1000                 7035
            numflock                       21                  137                 1000                 1000                    0
            numpty                          1                    3                  500                  500                    0
            numsiginfo                     53                  254                 1000                 1000                    0
            tcpsndbuf                 3144432             10843328              3145728              4194304          79847400259
            tcprcvbuf                  261776              5311712              2097152              3145728                    0
            othersockbuf                94376              4708440              2097152              3145728                 5084
            dgramrcvbuf                     0                16944             80530432            113298432                    0
            numothersock                   55                  626                 5000                 5000                    0
            dcachesize                4688185              7027826              9216000              9216000                    0
            numfile                      2142                 4118                24000                24000                    0
            dummy                           0                    0                    0                    0                    0
            dummy                           0                    0                    0                    0                    0
            dummy                           0                    0                    0                    0                    0
            numiptent                      38                  200                 1500                 1500                    0
Tue Sep  6 13:06:46 ICT 2011
~>
looking at the failcnt column, there are a lot of failures.
please increase barrier/limit, especially tcpsndnif at host first.
let's see if any improvement.
I have sent ticket to host, i will reply when barrier/limit and tcpsndnif has been tweaked. You can continue to monitor it as long as you want.

Thank you so much, NiteWave!
 
Last edited:

redstrike

Well-Known Member
#31
... The bottneck is the VPS provider limited my banwidth usage to 4 MB/s (normally my site used 6 MB/s), everything will be drop when it over the limit :(.

Thanks NiteWave, if you didn't find out that, i wouldn't know the situation. Adding more node for load balancing is my solution now.
 

NiteWave

Administrator
#32
while everything drop, CPU usage of litespeed process up to 100% -- so the reason is the bandwidth limit. this result will benefit other uesrs. Thanks for the feedback. I'd mark this thread "solved".
 
Top