Oninit cdi / Archive
Talk to an expert →

RE: Informix 7.23 connection problem

Posted in 2005

Summary

A user running Informix OnLine 7.22/7.23 on Unixware 2.1.3 found that once about 50 users connected, the server refused or dropped further connections, slowed down and eventually tied up the network card. Advice given: upgrade (not possible for them), check logs and OS/kernel parameters. NETTYPE was set to tlitcp,1,100,NET but changes to it appeared to be ignored at startup (the message log only reported LOCKS and BUFFERS changes), while other ONCONFIG changes took effect. A suggested workaround was adding a second sqlhosts entry/server alias on another port/service so extra users could connect via a different INFORMIXSERVER. No confirmed resolution is recorded; the thread ends asking about kernel parameter settings and release notes.

Auto-generated by DrWatson from the posts below — may be imperfect; read the full thread.

Topics: General Discussion

Simmons, Keith — — source: Usenet: comp.databases.informix

Cristoff 7.23 is very, very old. Any chance of upgrading to 7.31? Any log messages associated with these problems? Any O/S messages/errors being generated/ Keith -> -----Original Message----- -> From: Caitiff [mailto:viktor@customs.gov.mk] -> Sent: Monday, January 24, 2005 2:26 PM -> To: informix-list@iiug.org -> Subject: Informix 7.23 connection problem -> -> -> We are using Informix 7.23 Online database on Unixware -> 2.1.3. It works -> OK if less then 50 users are connected. When 51-st is connecting, it -> starts to refuse connections, create closed connections and -> go slow. I -> have increased tcp, MAXUP and NBUF parameters, but its not working. -> Cristoff -> ********************************************************************************** This message is sent in strict confidence for the addressee only. It may contain legally privileged information. The contents are not to be disclosed to anyone other than the addressee. Unauthorised recipients are requested to preserve this confidentiality and to advise the sender immediately of any error in transmission. This footnote also confirms that this email message has been swept for the presence of computer viruses, however we cannot guarantee that this message is free from such problems. ********************************************************************************** sending to informix-list

↪ replying to Simmons, Keith
Cristoff — — source: Usenet: comp.databases.informix

There is no chanse for upgrading. The system just close any next connection. After some time it blocks the network card. We have NETTYPE set up, but it probably needs initialisation.

↪ replying to Cristoff
Cristoff — — source: Usenet: comp.databases.informix

Our NETTYPE is set: NETTYPE tlitcp,1,100,NET NETTYPE ipcshm,1,250,CPU We did onmode -ky, oninit -y and onstat -m shows: INFORMIX-OnLine Version 7.22.UC2 -- On-Line -- Up 00:10:35 -- 38704 Kbytes Message Log File: /var/informix/online.log 11:04:01 Checkpoint Completed: duration was 1 seconds. 11:06:19 Checkpoint Completed: duration was 0 seconds. 11:06:20 INFORMIX-OnLine Stopped. Tue Jan 25 11:06:38 2005 11:06:38 Event alarms enabled. ALARMPROG = '/var/informix/log_full.sh' 11:06:38 DR: DRAUTO is 0 (Off) 11:06:39 INFORMIX-OnLine Initialized -- Shared Memory Initialized. 11:06:39 Physical Recovery Started. 11:06:40 Physical Recovery Complete: 0 Pages Restored. 11:06:41 Logical Recovery Started. 11:06:44 Logical Recovery Complete.0 Committed, 0 Rolled Back, 0 Open, 0 Bad Locks 11:06:44 Dataskip is now OFF for all dbspaces 11:06:44 On-Line Mode 11:06:45 Checkpoint Completed: duration was 0 seconds. 11:11:45 Checkpoint Completed: duration was 0 seconds. 11:16:45 Checkpoint Completed: duration was 0 seconds.

↪ replying to Cristoff
Valeriy Kravchuk — — source: Usenet: comp.databases.informix

> Our NETTYPE is set: > > NETTYPE tlitcp,1,100,NET > NETTYPE ipcshm,1,250,CPU Change these settings somehow (to NETTYPE tlitcp,1,60,NET for example), change some other parameter, like LOCKS (add 1000 more locks, it should not be very harmful), restart yor informix server, and send, please, the results of onstat -m after that. I also need to know the value of your ONCONFIG environment variable and the name of the file where you change the NETTYPE value. The results of the onstat -g ath after that and sqlhosts file's content may be of some use too. You know, I just don't believe that changes of NETTYPE in the _proper_ place are not taken into account on your platform. So, I try to figure out what happened. V.K.

↪ replying to Valeriy Kravchuk
Cristoff — — source: Usenet: comp.databases.informix

$ onstat -m INFORMIX-OnLine Version 7.22.UC2 -- On-Line -- Up 00:00:23 -- 39848 Kbytes Message Log File: /var/informix/online.log 13:12:11 Checkpoint Completed: duration was 0 seconds. 13:14:38 Checkpoint Completed: duration was 0 seconds. 13:14:39 INFORMIX-OnLine Stopped. Tue Jan 25 13:14:56 2005 13:14:56 Event alarms enabled. ALARMPROG = '/var/informix/log_full.sh' 13:14:57 DR: DRAUTO is 0 (Off) 13:14:58 INFORMIX-OnLine Initialized -- Shared Memory Initialized. 13:14:59 Physical Recovery Started. 13:14:59 Physical Recovery Complete: 0 Pages Restored. 13:14:59 Logical Recovery Started. 13:15:02 Logical Recovery Complete.0 Committed, 0 Rolled Back, 0 Open, 0 Bad Locks 13:15:02 Onconfig parameter LOCKS modified from 48000 to 50000. 13:15:02 Onconfig parameter BUFFERS modified from 12000 to 12500. 13:15:02 Dataskip is now OFF for all dbspaces 13:15:02 On-Line Mode 13:15:02 Checkpoint Completed: duration was 0 seconds. $ onstat -g ath INFORMIX-OnLine Version 7.22.UC2 -- On-Line -- Up 00:01:14 -- 39848 Kbytes Threads: tid tcb rstcb prty status vp-class name 2 bfb2188 0 2 sleeping(Forever) 3lio lio vp 0 3 bfb240c 0 2 sleeping(Forever) 4pio pio vp 0 4 bfb2690 0 2 sleeping(Forever) 5aio aio vp 0 5 bfb2914 0 2 sleeping(Forever) 6msc msc vp 0 6 bfb3110 0 2 sleeping(Forever) 7aio aio vp 1 7 bfb3400 bf54014 4 sleeping(secs: 1) 1cpu main_loop() 8 bfb3ae0 0 2 running 1cpu sm_poll 9 c02fbf0 0 2 running 8tli tlitcppoll 10 c03f148 0 2 sleeping(Forever) 1cpu sm_listen 11 c03f6ac 0 2 sleeping(secs: 2) 1cpu sm_discon 12 c03fab4 0 3 sleeping(Forever) 1cpu tlitcplst 13 c04b0c0 bf5444c 2 sleeping(Forever) 1cpu flush_sub(0) 14 c05d860 0 2 sleeping(Forever) 9aio aio vp 2 15 c05db50 0 2 sleeping(Forever) 10aio aio vp 3 16 c05dd5c 0 2 sleeping(Forever) 11aio aio vp 4 17 c07f028 0 2 sleeping(Forever) 12aio aio vp 5 18 c07f18c 0 2 sleeping(Forever) 13aio aio vp 6 19 c07f398 0 2 sleeping(Forever) 14aio aio vp 7 20 c07f5a4 0 2 sleeping(Forever) 15aio aio vp 8 21 c07f7b0 0 2 sleeping(Forever) 16aio aio vp 9 22 c07f9bc 0 2 sleeping(Forever) 17aio aio vp 10 23 c07fbc8 0 2 sleeping(Forever) 18aio aio vp 11 24 c07fdd4 0 2 sleeping(Forever) 19aio aio vp 12 25 c08b028 0 2 sleeping(Forever) 20aio aio vp 13 26 c08b214 0 2 sleeping(Forever) 21aio aio vp 14 27 c08b420 0 2 sleeping(Forever) 22aio aio vp 15 28 c08b62c 0 2 sleeping(Forever) 23aio aio vp 16 29 c08b838 0 2 sleeping(Forever) 24aio aio vp 17 30 c08ba44 0 2 sleeping(Forever) 25aio aio vp 18 31 c08bc50 0 2 sleeping(Forever) 26aio aio vp 19 32 c03fd58 bf54884 2 sleeping(secs: 49) 1cpu btclean 48 c0c5524 bf5552c 4 sleeping(secs: 1) 1cpu onmode_mon 54 c2050c0 bf54cbc 2 cond wait(sm_read) 1cpu sqlexec 55 c0b1028 bf550f4 2 cond wait(sm_read) 1cpu sqlexec 56 c06b1f4 bf55964 2 cond wait(sm_read) 1cpu sqlexec 58 c09b074 bf55d9c 2 cond wait(sm_read) 1cpu sqlexec 60 c1011e0 bf561d4 2 cond wait(sm_read) 1cpu sqlexec We have the same ONCONFIG value and onconfig_file_name: onconfig.server1 sqlhosts_file_values: server1 onipcshm server1 onlines server1_i ontlitcp server1 online

↪ replying to Cristoff
Valeriy Kravchuk — — source: Usenet: comp.databases.informix

> INFORMIX-OnLine Version 7.22.UC2 -- On-Line -- Up 00:00:23 -- 39848 > Kbytes > > Message Log File: /var/informix/online.log > 13:12:11 Checkpoint Completed: duration was 0 seconds. > 13:14:38 Checkpoint Completed: duration was 0 seconds. > 13:14:39 INFORMIX-OnLine Stopped.> > Tue Jan 25 13:14:56 2005 > > 13:14:56 Event alarms enabled. ALARMPROG = > '/var/informix/log_full.sh' > 13:14:57 DR: DRAUTO is 0 (Off) > 13:14:58 INFORMIX-OnLine Initialized -- Shared Memory Initialized. > 13:14:59 Physical Recovery Started. > 13:14:59 Physical Recovery Complete: 0 Pages Restored. > 13:14:59 Logical Recovery Started. > 13:15:02 Logical Recovery Complete.> 0 Committed, 0 Rolled Back, 0 Open, 0 Bad Locks > > 13:15:02 Onconfig parameter LOCKS modified from 48000 to 50000. > 13:15:02 Onconfig parameter BUFFERS modified from 12000 to 12500. > 13:15:02 Dataskip is now OFF for all dbspaces > 13:15:02 On-Line Mode > 13:15:02 Checkpoint Completed: duration was 0 seconds. So, you had changed 3 parameters, including NETTYPE, and only 2 of them are really taken into account during server startup, right? I have not seen this kind of behavior on any UNIX platform before... Please, read the release notes for your platform carefully ($INFORMIXDIR/release/en_us/*/*). May be, it is known bug or limitation for your release. The only platform, to my knowledge, where NETTYPE settings are ignored, is Windows... > server1 onipcshm server1 onlines > server1_i ontlitcp server1 online Possible workaround is to create yet another server alias, pointing to the third row in the sqlhosts, that allows to work using ontlitcp, but through a different service (not online), or port. Then 50 more users may be able to connect, but to the different INFORMIXSERVER. I do not know how can I really help you, if NETTYPE is ignored. Unbelievable. Sorry. May be, somebody else... Last attempt: had you set proper values of all the kernel parameters mentioned in the release notes? Please, list them all. V.K.

Related threads