Petty comments on Version 15 ...
Posted in 2026
Jon Ritson listed minor Informix 15.0.1.0DE oddities: a duplicated USERMAPPING onconfig warning, a KAIO sector size warning only on first init, and no FC in the version name. Andreas Legner could not reproduce the first two; the third was a naming change. Vicente Salvador Cubedo explained 4096 sectors; Andreas also showed v15 lifts the 32767 output rowsize limit.
Auto-generated by Claude from the posts below — may be imperfect; read the full thread.
Topics: Performance & Tuning, Storage & Space Management, Server Administration
Hi ... just because ...
1.
Warning: Parameter duplicated in config file (/products/15.01.FC1/etc/onconfig_leap16_j_a).
Only the first occurrence will be used. (USERMAPPING)
> grep USERMAPPING onconfig.std
# USERMAPPING - Control access to the server for users without operating
USERMAPPING OFF# USERMAPPING - Control access to the server for users without operating
USERMAPPING OFF2. Initialising with DIRECT_IO set to 1 and the following is reported
14:29:04 IBM Informix Dynamic Server Version 15.0.1.0DE
14:29:04 Performance Advisory: Currently IBM Informix Dynamic Server cannotsupport KAIO on the device containing
'/instances/leap16_j_a/chunks/rootdbs'
Its Sector Size (4096) and the page size (2048) are incompatible.
14:29:04 Results: Direct and concurrent I/O are disabled for this chunk.
14:29:04 Action: Standard AIO (possibly buffered) will be used for I/Oto this chunk.
However, after just a restart of informix and no other change, kernel aio is all good
> onstat -g ioq
IBM Informix Dynamic Server Version 15.0.1.0DE -- On-Line -- Up 00:14:10 -- 563264 Kbytes
2026-04-08 14:44:00 -- Infrastructure Version: 1
AIO I/O queues:q name/id len maxlen totalops dskread dskwrite dskcopy
fifo 0 0 0 0 0 0 0
drda_dbg 0 0 0 0 0 0 0
sqli_dbg 0 0 0 0 0 0 0
kio 0 0 32 4538 1667 2871 0
kio 1 0 32 1870 720 1150 0
kio 2 0 32 3191 860 2331 0
kio 3 0 33 1263 917 346 0
...
3. Version 15.01.0DE
informix@leap16-1-a:/products/15.01.FC1/etc> oninit -version
Program Name: oninit
Build Version: 15.0.1.0DE
Has naming convention changed (i.e. no "FC" like ... 15.01.FC0??)?
------------------------------
Jon Ritson
------------------------------
Hi JJ, not seeing your 1., in none of my 6 or so v15 installations. not seeing your 2. either, after initializing (-ivy) using a 4k sector size disk for my ROOTPATH (and DIRECT_IO 1). I do see KAIO being added (queues, threads) once defining more chunks on other disks. And yes to your 3. -> Fix pack naming conventions Cheers, Andreas ------------------------------ Andreas Legner Informix Dev HCL Software ------------------------------
Thanks for the response Andreas ... "moving on" :)
informix@leap16-1-a:/products/15.01.FC1/etc> oninit -version
Program Name: oninit
Build Version: 15.0.1.0DE
Build Number: 147
Build Host: lxv15nb1
Build OS: Linux kernel-4.18.0-477.10.1.el8_8.x86_64 glibc-2.28-251.el8_10.25.i686 glibc-2.28-251.el8_10.25.x86_64
Build Date: Tue Dec 16 06:44:25 CST 2025
Build Timestamp: 2025-12-09T13:46:56-06:00
Build Commit: ea65e6c33153a1cc52f0f3286fdf7e49bacee28a
GLS Version: glslib-8.0.0.2
informix@leap16-1-a:/products/15.01.FC1/etc> oninit -V
IBM Informix Dynamic Server Version 15.0.1.0DE
Tue Dec 16 06:44:25 CST 2025
------------------------------
Jon Ritson
------------------------------
About kayo message this behavior is expected. You have a disk sectors physically formated with 4096 bytes. Only pages with size multiple of this sector size can implement KAIO. To fix this, you can reformat your disk with 512 bytes. Then all page sizes will be multiple of 512 and KAIO will be enabled in all page sizes. Vicente Salvador ------------------------------ Vicente Salvador Cubedo Software Architect DEISTER, S.A. Barcelona +34 932063298 ------------------------------
Just an Informix restart and no error is reported and kaio is fine - only appears on oninit -i.
openSUSE 16 on btrfs
Sent from Sky Yahoo Mail for iPhone
leap16-1-a:~ # lsblk -t NAME ALIGNMENT MIN-IO OPT-IO PHY-SEC LOG-SEC ROTA SCHED RQ-SIZE RA WSAME sda 0 512 0 512 512 0 mq-deadline 64 1024 0B ├─sda1 0 512 0 512 512 0 mq-deadline 64 1024 0B ├─sda2 0 512 0 512 512 0 mq-deadline 64 1024 0B └─sda3 0 512 0 512 512 0 mq-deadline 64 1024 0B sr0 0 2048 0 2048 2048 0 mq-deadline 2 1024 0B leap16-1-a:~ # ls -l /instances/leap16_1_a/chunks/rootdbs -rw-rw----. 1 informix informix 268435456 Apr 9 12:01 /instances/leap16_1_a/chunks/rootdbs leap16-1-a:~ # mount | grep " / " /dev/sda2 on / type btrfs (rw,relatime,seclabel,ssd,space_cache=v2,subvolid=266,subvol=/@/.snapshots/1/snapshot) leap16-1-a:~ # df -h /instances/leap16_1_a/chunks/rootdbs Filesystem Size Used Avail Use% Mounted on /dev/sda2 62G 16G 46G 26% / ------------------------------ Jon Ritson ------------------------------
So you're saying your ROOTPATH is on /dev/sda2 which resides on a 512-bytes sector size device, and still oninit -iy would report "
Sector Size (4096)", but subsequent restarts wouldn't?
I just tried the same, abusing our machine's /dev/sda2 (mounted as /) with very similar lsblk -t output -> no problem.
------------------------------
Andreas Legner
Informix Dev
HCL Software
------------------------------
Yes ... openSUSE 16 (Leap) VM s ... UNDERLYING disk at the HOST is: [root@jonnytdi utils]# mount | grep cd /dev/sdc2 on /cd_caddy_ssd type ext4 (rw,noatime,seclabel,discard) [root@jonnytdi utils]# lsblk -t NAME ALIGNMENT MIN-IO OPT-IO PHY-SEC LOG-SEC ROTA SCHED RQ-SIZE RA WSAME ... sdc 0 4096 0 4096 512 1 mq-deadline 64 128 0B ├─sdc1 3072 4096 0 4096 512 1 mq-deadline 64 128 0B └─sdc2 0 4096 0 4096 512 1 mq-deadline 64 128 0B And the VM Virtual Disk is presented as ... /cd_caddy_ssd/virtual_machines/Leap_16/Leap16/leap16-1-a/leap16-1-a.vdi So it looks like the LOGICAL sector size IS 512 but the PHYSICAL is 4096 at the host. S'all good mate ... not really worried as "it works for me" (just rebuilt 8 instances and converted them into two clusters in under an hour so I'm 'appy and far quicker than USB3 hanging off the host ... even got HUGE PAGES enabled at the VM level :O ) ------------------------------ Jon Ritson ------------------------------
Hello community, I apologize in advance for being lazy and asking this question here instead of checking/testing myself (embarrassed to say I have not setup/tested Informix 15 yet – insisting on keeping running the nicely working 12 and some 14 installations for our various sites/systems). Does version 15 address this annoying error we get when we have many large CHAR, LVARCHAR etc. columns in a query: Maximum output rowsize (32767) exceeded Thank you, Hal Maner
It does. Try this: select owner::char(32767),tabname::char(32767) from systables where tabid=1; And this is only the output rowsize (tuple size) - wait until also the (same) table rowsize limit will fall ------------------------------ Andreas Legner Informix Dev HCL Software ------------------------------
Related threads
- Posting from the Informix-list
- Migrating from IDS 9.40.UC6 to 11.50.UC3
- Ip for a network session
- questions onstat -g