ap_6200 Posted February 18, 2008 Report Share Posted February 18, 2008 One other thing that would be really nice on the voicemail system is if they could add "hangup detection" and not issue a voicemail if that is all it is. We get a lot of voicemails that are just hangups, and it wastes a lot of time. There have to be some common voice processing functions that can determine that pretty easily, especially when the message is less than 10 seconds. Thanks Quote Link to comment Share on other sites More sharing options...
shopcomputer Posted February 19, 2008 Report Share Posted February 19, 2008 I also got several complaints about blank / hang up messages. I also got a few that voicemail volume is very low. Quote Link to comment Share on other sites More sharing options...
ap_6200 Posted February 19, 2008 Author Report Share Posted February 19, 2008 I also got several complaints about blank / hang up messages. I also got a few that voicemail volume is very low. I echo the volume low issue as well - I've reported it before, but was discounted on a few levels - so I let it go. I get fewer of those than I get "my phones are saying URL call disabled" because my pbxnsip service decided to shut down with no error indication" Quote Link to comment Share on other sites More sharing options...
shopcomputer Posted February 19, 2008 Report Share Posted February 19, 2008 I always set service to auto restart. I did that the first time the service crashed. I am not sure if I ever had a service crash since, as it would start right back up, before the complaints come in. Quote Link to comment Share on other sites More sharing options...
ap_6200 Posted February 19, 2008 Author Report Share Posted February 19, 2008 I always set service to auto restart. I did that the first time the service crashed. I am not sure if I ever had a service crash since, as it would start right back up, before the complaints come in. I would rather it not crash - but i get the point - I will probably start it back doing that tomorrow morning. Quote Link to comment Share on other sites More sharing options...
andrewgroup Posted February 19, 2008 Report Share Posted February 19, 2008 I would rather it not crash - but i get the point - I will probably start it back doing that tomorrow morning. IMHO, Voice servers are no different than other servers, and in some ways even more critical. I could schedule a LAN Server upgrade during client work hours with some advanced notification, but not likely could we shut down their telephone system. We have avoided system crashes by investing in server class systems. Lowend Dell Poweredge, 400/800 series, SuperMicro, All with ECC RAM, HARDWARE RAID, and a very strict OS installation setup. Our oldest Windows and Linux machines are going on 2 years back to version 1.5.x with no real crash experiences. Quote Link to comment Share on other sites More sharing options...
ap_6200 Posted February 20, 2008 Author Report Share Posted February 20, 2008 IMHO, Voice servers are no different than other servers, and in some ways even more critical. I could schedule a LAN Server upgrade during client work hours with some advanced notification, but not likely could we shut down their telephone system. We have avoided system crashes by investing in server class systems. Lowend Dell Poweredge, 400/800 series, SuperMicro, All with ECC RAM, HARDWARE RAID, and a very strict OS installation setup. Our oldest Windows and Linux machines are going on 2 years back to version 1.5.x with no real crash experiences. Put them on 2.1.x and put it on the server of your choice and let me know if they make it 30 days without restarting - and have them do about a 7K call volume per day. One domain is doing 4K by itself. The "server" doesn't crash - the application does. Quote Link to comment Share on other sites More sharing options...
andrewgroup Posted February 20, 2008 Report Share Posted February 20, 2008 Put them on 2.1.x and put it on the server of your choice and let me know if they make it 30 days without restarting - and have them do about a 7K call volume per day. One domain is doing 4K by itself. The "server" doesn't crash - the application does. Our experience is different. So far we are over 3 months on 2.1.5.2357. For the most part we keep a 23 Channel PRI, 70% utilized 5 business days per week with ACT of 2 minutes, thus about 3,800 calls per day with 45 internal SNOM 320 phones. This has been the pace since we upgraded to release 2.0.3 and now on 2.1.5.2357, I think the largest month was 65,000 calls. 3 Nics one to the Local LAN, 1 to the PRI Gateway, and another to what will soon be 6 remote offices with pbxnsip. I placed some documents on this system in the best practices on our windows Installations. Do these agree with what you are doing? We also have a very standard Gentoo install from source that we can use where needed. I've included both Free Virtual RAM and CPU performance charts over the last 3 months. The single instance of the RAM following was us remotely accessing the system and doing some serious computing to extract a bunch of files...The system returned with more available Virtual RAM. It's a Windows thing, it must have cleaned and release the pool when it dipped low for our remote process. Do you have any Windows WMI, SNMP, or other agents monitoring the system to indicate the time/day/load where the problem is occurring? it would be interesting to triangulate when the problem occurs. Perhaps it's event driven and not load driven. we simply see little to no load on our systems. We also fully track all traffic, events on all switches, routers, using our NMS, and SolarWinds. The Snoms are still using 6.5.x firmware and we are running SRTP internally.. (despite claims of others having troubles with 6.5.x and SRTP) Perhaps someone can step up with a SIP session tester from IXIA and we can fully load one of these puppys's to the MAX... It would be nice to see where the smoke comes from...At least we'd know the upper limit, and everything has an upper limit. Quote Link to comment Share on other sites More sharing options...
ap_6200 Posted February 20, 2008 Author Report Share Posted February 20, 2008 Our experience is different. So far we are over 3 months on 2.1.5.2357. For the most part we keep a 23 Channel PRI, 70% utilized 5 business days per week with ACT of 2 minutes, thus about 3,800 calls per day with 45 internal SNOM 320 phones. This has been the pace since we upgraded to release 2.0.3 and now on 2.1.5.2357, I think the largest month was 65,000 calls. 3 Nics one to the Local LAN, 1 to the PRI Gateway, and another to what will soon be 6 remote offices with pbxnsip. I placed some documents on this system in the best practices on our windows Installations. Do these agree with what you are doing? We also have a very standard Gentoo install from source that we can use where needed. I've included both Free Virtual RAM and CPU performance charts over the last 3 months. The single instance of the RAM following was us remotely accessing the system and doing some serious computing to extract a bunch of files...The system returned with more available Virtual RAM. It's a Windows thing, it must have cleaned and release the pool when it dipped low for our remote process. Do you have any Windows WMI, SNMP, or other agents monitoring the system to indicate the time/day/load where the problem is occurring? it would be interesting to triangulate when the problem occurs. Perhaps it's event driven and not load driven. we simply see little to no load on our systems. We also fully track all traffic, events on all switches, routers, using our NMS, and SolarWinds. The Snoms are still using 6.5.x firmware and we are running SRTP internally.. (despite claims of others having troubles with 6.5.x and SRTP) Perhaps someone can step up with a SIP session tester from IXIA and we can fully load one of these puppys's to the MAX... It would be nice to see where the smoke comes from...At least we'd know the upper limit, and everything has an upper limit. This is what we experience on a daily basis... Version: 2.1.5.2357 (Win32) License Status: License Duration: Permanent Additional license information: Extensions: 224/9999 Accounts: 283/9999 Working Directory: C:\Program Files\pbxnsip\PBX IP Addresses: 127.0.0.1 10.1.0.10 x.x.x.x (public) MAC Addresses: Calls: 5621/2113 (CDR: 5355) 0/0 Calls SIP packet statistics: Tx: 873362 Rx: 865318 Emails: Successful sent: 107 Unsuccessful attempts: 0 Uptime: 1 10:05:37 (73MB/2047MB 22% 40666880-0) WAV cache: 0 CPU_UTIL.bmp Quote Link to comment Share on other sites More sharing options...
andrewgroup Posted February 20, 2008 Report Share Posted February 20, 2008 This is what we experience on a daily basis... You Win. Did you see my post - pbxnsip challenge http://forum.pbxnsip.com/index.php?showtopic=717 I would fully expect any SIP system will have an upper limit and knowing when that occurs would be a reasonable thing to know. (pbxnsip compared to a Cisco BTS 10200 - Catalyst Investment) is still such a good trade. I revisited my test case and 20 days ago the PBX was powered off during maintenance. The company is a seasonal contractor and the cold has all but stopped all business, so our peak call periods won't occur until april thru Oct. Still a 20 day count for this period shows the system is quite busy. I'll tag this system to avoid doing a restart or maintenance for as long as we can or until we have to. Calls: 19370/11392 (CDR: 8057) 0/0 Calls SIP packet statistics: Tx: 22979436 Rx: 22980262 Emails: Successful sent: 171 Unsuccessful attempts: 0 Uptime: 20 06:47:12 (49MB/511MB 77% -1609395252--1631031348) WAV cache: 0 are you managing this with S11? Are you on SRVR2k3 WinXP, Linux? Are you collecting WMI data across the entire server? How About SNMP or Syslog data on all infrastucture? The reason I say, is that if you are resetting the service to get it started again, I clearly reach your call counts Quote Link to comment Share on other sites More sharing options...
andrewgroup Posted February 20, 2008 Report Share Posted February 20, 2008 More 411, If you reduce logging to a minimum (disable) does this extend the running period? I would install a usable WMI tool to collect and report on all WMI available points. paessler has a free 30 day trial.. We own N-Central from Nable for managing our Servers. I could give you an agent to run for a test period. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.