xaminmo: Josh 2016 (Default)
If I remove the first main adapter, and re-add it, then I can add/remove either adapter or IP interface after that.

If I remove the second main adapter, and re-add it, then I cannot remove the first, and dropping the IP interface crashes.

So, assuming adapter_names=ent2,ent6

This works
Read more... )
xaminmo: Josh 2016 (Default)
computersarefun put in a request for AIX to auto-grow/shrink filesystems.
Ref: https://www.ibm.com/developerworks/rfe/execute?use_case=viewRfe&CR_ID=114789

This seems more like a monitoring thing than an operating system thing.
Also, handling this as a thin LUN is probably better where
Read more... )
xaminmo: Josh 2016 (Default)
JFS2 Unsupported
Putting JFS2 on non-LVM block devices has been working for a long time. I​ wrote up how to put JFS2 on a ramdisk back at AIX 4.3.3. I lost the techdoc from back then, but IBM has a newer re-write dated 2008 here: http://www-01.ibm.com/support/docview.wss?uid=isg3T1010722

JFS2 requires the underlying system to tell it if something goes away, or for it to stay there as long as the filesystem is mounted. LVM does this for disk, and the ramdisk drivers do this as well (mostly because if the ramdisk fails, likely the system has failed). The key there is that with JFS2, the ramdisk pages are pinned.

I wrote up including performance on USB 1 and USB2 ports in January of 2010 HOWTO: JFS2 on USB device on AIX Everything seems fine, and dandy, even mount on boot, except it's not supported by AIX Development.

JFS2 Problems
The problem for USB Mass Storage Devices is that the device can just go away unexpectedly. If a disk goes into deep sleep, or resets because of a loose connection, the JFS drivers do not get notified. So, they take writes, and JFS2 saves them up until it's time to flush. It goes to flush, and the I/O channel is gone. Sometimes, this is just loss of everything in cache. Sometimes, config methods hang until reboot. Other times, the system crashes.

​Because of that, we still cannot put LVM on a USB Mass Storage Device. This would take changes to notification of device availability, perhaps changes to the sync daemon, etc. Who knows, but there's not been enough push from paying customers to make it a priority for AIX Development. Until that happens, don't expect formal support for JFS2 on these devices.

UDF is the solution
AIX development supports read/write and even booting from USB Mass Storage Devices, but only with UDFS. The purpose is for writing a mksysb (system boot) image, or tar/cpio files, etc, and exists because of the RDX USB Internal Dock sold with some systems.

​Boot support is provided as well: REF: ​http://www-01.ibm.com/support/docview.wss?uid=isg1IZ66737

More info on RDX USB Internal Dock. https://www.ibm.com/support/knowledgecenter/POWER7/p7hdt/fc1103.htm

However, RDX is just a hot-swap USB to SATA drive bay. Any current USB drive (USB3 is preferred due to performance), should work fine.

HOWTO: Create, Read, and Write UDF on AIX

Create a bootable filesystem
mksysb -eXpi /dev/usbms0

Create an empty filesystem
udfcreate -d /dev/usbms0

Create a UDF 2.01 filesystem
udfcreate -f3 -d/dev/usbms0

NOTE: UDF 2.01 supports a real-time filesystem. It's still UDF, so don't try to put a database, or a million files on there.

Access read/write
mount -vudfs /dev/usbms0 /USBDRIVE

NOTE: The mksysb is a SPOT, plus a mksysb image, so adding files to the UDF will not make the restore huge.

USB Adapters on AIX
Add-in USB3 XHCI adapter for POWER8 is:
* CCIN 58F9 - PCIE2 4-port USB3 adapter
* FC EC45 and FRU 00E2932 for Low Profile
* FC EC46 and FRU 00E2934 for full height.
* driver is 4c1041821410b204 internal or 4c10418214109e04 PCIe

Add-in USB2 EHCI adapter for POWER7 is:
* CCIN 57D1 - PCI-E 4-port USB2 adapter
* driver is 33103500 integrated or 3310e000 PCIe
* FC 2728 or FRU 46K7394

Add-in USB2 EHCI adapter for POWER6/POWER5 is:
* CCIN 28EF - PCI 2-port USB2 adapter
* FC 2738 or FRU 80P2994
* Belkin F5U219 - exact same card without the sticker.
* driver is 99172604 internal or 99172704 PCI

Original USB1 OHCI /UHCI adapter for POWER5 and earlier was
* driver 22106474 on blades or c1110358 PCI
* This device is not really available anymore.

Original post:
xaminmo: Josh 2016 (Default)
This changes periodically, but for today, here is what I would do.

My PowerHA selection process would be:
* 7.1.3 SP06 if I needed to deploy quickly, because I have build docs for that.
* 7.1.4 doesn't exist, but if it came out before deployment, I would consider it. Whichever was a newer
Read more... )
xaminmo: Josh 2016 (Default)
QEMU on Windows will run ppc64 and ppc64le emulation.
It emulates the same as what PowerKVM on an S812L would provide.
It's kind of slow because there is no KVM module, AND Intel vs PPC,
AND emulator mode is single-core/proc/thread.

You can get Windows installer
Read more... )
xaminmo: Josh 2016 (Default)
PROBLEM: On some migrates, we found the rpdomain would not stay running on one node.
The cluster was up, and SEEMED to operate normally, but errpt got CONFIGRM stop/start messages every minute.

lsrpdomain would show Offline, or "Pending online".

lsrpnode would show:
2610-412 A Resource
Read more... )
xaminmo: Josh 2016 (Default)
This is why powerpath for boot devices is a BAD thing. At some point, someone will put a non-powerpath device with a powerpath device inside of rootvg on a production server. Then you end up completely broken:

root@somehost:/>bosboot -ad /dev/hdiskpower7
0301-154 bosboot: missing proto file: /usr/lib/boot/network/chrp.hdiskpower.proto

### This error means you need to run "pprootdev fix"

root@somehost:/>pprootdev fix
pprootdev: PowerPath boot is not currently enabled.

root@somehost:/>lspv | grep rootvg
hdiskpower7 FFFFFFFFBBBBBBBB rootvg active

root@somehost:/>pprootdev on
bosboot verification failed.
Run 'bosboot -vd /dev/ipldevice' to determine cause of failure.

root@somehost:/>bosboot -vd /dev/ipldevice
0301-154 bosboot: missing proto file: /usr/lib/boot/network/chrp.hdiskpower.proto

### I tried to reinstall powerpath, and that patently failed. I need to bring the apps offline, then remove powerpath0, then reboot, then remove powerpath0, then uninstall, then reinstall, then configure powerpath, then reboot, then pprootdev on, then reboot, then pprootdev fix.
That's assuming it will all actually work.

### Sigh. Time to shave a yak.

Or really, time to move to MPIO because they're running reserve_policy=single_path anyway.
xaminmo: (Logo IBM CATE)
I always run into issues when I work in a multiple VLAN environment, because it's not *that* common for my builds. This is a reminder for me.

The magic is when using multiple VLANs:
1) Don't use the real VLAN ID for the trunk PVID unless you know for certain that was set on the switch. It is stripped off of all packets, and who knows what the PVID of the switch is, if any.
2) Any mismatch between PVID on the SEA and the trunk will cause packets to be dropped.
3) Don't use IEEE VLAN mode for the client adapter unless you're going to add VLAN interfaces from AIX. When not in VLAN mode, the PVID is ADDED to all packets on client adapters.
4) When using multiple trunks on one SEA, they all have to be the same trunk priority. ha_mode=sharing balances not using trunk priority, but based on the order of the virt_adapters field.
xaminmo: Josh 2016 (Default)
Was installed on the single, internal SAS disk, but was provisioned some 60G LUNs from a VNX (aka CLARiiON).
Got the MPIO drivers on (because powerpath boot has management encumberances), and verified the devices looked good.
Rebooted, the mirrored onto the new LUNs, removed the old SAS disk, updates bootlist and bosboot.
shutdown -Fr and I get this:

QLogic QMI3572  Host Adapter FCode(IBM): 3.14 2010-04-30 15:03
ISP Firmware version 5.03.06
Elapsed time since release of system processors: 37167 mins 43 secs
DEFAULT CATCH!, exception-handler=fff00400
at   %SRR0: 0000000048000104   %SRR1: 0000000040003002
Open Firmware exception handler entered from non-OF code

Client's Fix Pt Regs:
 00 0000000000000001 0000000000038160 0000000060000000 0000000000044bd0
 04 0000000000000001 0000000000000000 0000000000000100 0000000000044bd0
 08 ffffffffffffffff 0000000000000003 0000000048000104 0000000060000000
 0c 0000000000000078 0000000000000000 0000000000000000 0000000000000000
 10 0000000000000000 0000000000000000 0000000000000000 0000000000000000
 14 0000000002000000 0000000000000008 0000000000000000 0000000000000000
 18 0000000000000000 0000000000000001 0000000000039b58 0000000000044e10
 1c 0000000000044e08 0000000000038350 0000000000000000 0000000000000000
Special Regs:
    %IV: 00000400     %CR: 22000040    %XER: 00000000  %DSISR: 00000000
  %SRR0: 0000000048000104   %SRR1: 0000000040003002
    %LR: 0000000000011d7c    %CTR: 0000000048000104
   %DAR: 0000000000000000
Virtual PID = 0
0 > 

It's been years since I got a DEFAULT CATCH so I don't remember where to go from here.

Google is giving me no joy.
xaminmo: Josh 2016 (Default)
Not sure where to look for this. This problem exists even after cloning the system via mksysb.
lsattr -El inet0, or en0, or en1, or lo0 shows each attribute duplicated.
odmget for CuDv and CuAt do not show duplicate lines.
lsattr -El on fscsi0, hdisk0, sys0, mem0, proc0 and such are all normal.
Google doesn't help - it just shows entries about duplicate routes, which is not the problem here.

Any ideas?

Command output details )
xaminmo: (Logo IBM AIX 3.2.5)
So far, IBM says you cannot mount a USB mass storage device. I wanted to test this out a bit.
Read more... )
Suffice it to say that if you're willing to do a little but of manual work for the set-up, it's very usable.


Nov. 23rd, 2007 04:11 pm
xaminmo: (Gloomy Angry)
So, a cron job fails on AIX 5300-05-06.

The machine gets rebooted to enable aio.

From then on, cron no longer processes the crontab for the user whose job failed.

WTF is THAT about?

I had to edit the cron tab and save it changed to get it to continue.

This is *not* acceptable.
xaminmo: Josh 2016 (Default)
ssh hscroot@hmcname
for i in `lssyscfg -r sys -F name` ; do chled -r sa -t phys -o off -m $i ; done
xaminmo: Josh 2016 (Default)
Ping was running from the client to a different client.
I would pull one cable, replug, pull the other.
No provisions to test link up but no traffic.

The longest delay was when I did the second pull very quickly
after replugging the first cable, and this was still less than 1 second
in both instances.

The ping provision was not tested due to no access to VLAN the port(s).
Also, if the GW goes down, I don't know how NIB and SEA-ha handle this.
Would there be loss of network due to flap, or would failover just abort?
environment and setup )

Flash Copy

Jan. 17th, 2007 03:17 pm
xaminmo: Josh 2016 (Default)
Since I had so much trouble finding any clear, simple, concise document containing this info, I figure it's best to post it out here for everyone else.

This hasn't been tested, and likely contains some manner of typos, but this should give you the jist of it.
AIX JFS2 snapshotting )DS4000 SMcli flashcopy )
xaminmo: Josh 2016 (IBM AIX Advanced eXecutive Interactive)
I spent 4 hours on this. It took me 2 hours to decide to do this rather than manually create.
Then it took one hour to go from a command, to a loop, to a loop of loops, to this stuff.
Most of that time was spent trying to figure out how to get the environment variable into awk.
I had to export it in KSH, and I had to do the wierd ENVIRON assignment. -v and other things didn't help.
Did I mention, I don't know awk? ok.

After that, I had to resolve failures, which amounted to laziness in picking the jfslog - ok, fixed;
and also in assuming "jfs" volumes were filesystems. I changed it to grep out "N/A".

These steps resolved all but 1 out of 659 logical volumes. The last of which was an exception to my chosen homogenization of the filesystem parameters. nbpi of 8192 was too small for that big of a logical volume. Aparently, JFS1 has an upward limit to the number of inodes that may be in a filesystem. Should be ok. It's all DB containers anyway. I manually re-ran this one crfs.

I also manually ran the mkvg commands. I didn't feel like making a BIG tool out of this.

So anyway, This is a fairly reusable bit of code for replicating the VG structure of one box onto another.

It's not a script to be run by itself. It doesn't pick up all of the VG, LV or FS parameters. This was on AIX 4.3.3, but should be portable. This assumes alot, but it worked for me.

Taylor to suit your needs. If you break it, you buy it.
Read more... )
xaminmo: Josh 2016 (Default)
From current AIX media, base install, they give this error:
  Filesets listed in this section failed pre-installation verification
  and will not be installed.

  Missing Filesets
  The following filesets could not be found on the installation media.
  If you feel these filesets really are on the media, check for typographical
  errors in the name specified or, if installing from directory, check for
  discrepancies between the Table of Contents file (.toc) and the images that
  reside in the directory.


  << End of Failure Section >>

I think these have been superceded, but I'm not sure.

It looks messy though.

The fun thing is that base install really just restores a tiny mksysb, then installs a buttload of filesets.

xaminmo: Josh 2016 (Default)
OK, so a fully decked out p595 can also have 60 xSeries 2gHz Pentium M servers inside, as long as you're running at least one AS/400 partition. So, technically, you could have 124 processors, all at about 2GHz, all in one or two frames.

IXS for iSeries

When you add these, they don't have disks. You use the AS/400 command ADDNWSSTGL (Add Network Server SToraGe Link) to link a virtual disk from AS/400 to the Network Server (aka the Integrated X(series|86) Server) PCI-X 2-slot card, and then you can use os/400 to lay down an image for you onto that disk (SuSE, RedHat, Win2k, Win2k3, etc).

The only drawback is that you ALSO need an I/O Processor card. This goes on the same PCI bus as the IXS card and OS/400 sends all I/O requests and DMA through that card so the OS and system processors don't have to sit waiting for a response.

Still, that's a pretty good technogy. Costly, but it stretches your CPU power having coprocessors for all of your I/O slots.


xaminmo: Josh 2016 (Default)

July 2018

12 34567
15 161718192021


RSS Atom

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated Apr. 26th, 2019 01:58 am
Powered by Dreamwidth Studios