?

Log in

No account? Create an account

Josh-D. S. Davis

Xaminmo / Omnimax / Max Omni / Mad Scientist / Midnight Shadow / Radiation Master

Previous Entry Share Next Entry
Complaint
Josh 201604 KWP
joshdavis
Follow with me here.

Definitions
LPP Source = repository of installable filesets
SPOT = NFS exported /usr
mksysb = backup format image of everything in the root volume group
ODM = AIX database used to store device/config information.

Restore internal procedure
1. network boot and mount up the SPOT
2. Build the root volume group and mount the filesystems
3. restore the mksysb into the new root filesystem mounted on /mnt
4. overmount the new root filesystem to /
. NOTE: At this point, you are running the kernel from the boot server but the kernel in /unix is from the mksysb.
5. Install any missing device drivers from the lppsource
NOTE: package dependancies must be met already due to the risk of otherwise having to update every fileset, some of which may not be available.
6. boot image is re-made
7. system is rebooted.

Common problems
If the mksysb and SPOT don't match, the overmount of rootvg may work, but then lots of libc.a linker failures will occurr when scripts and commands try to run.

If the lppsource doesn't match, the drivers aren't installed and the system fails to boot wit LED 554 (can't find the root disk) or 512/518 (can't mount / or /usr).

Event Log
ok, so, the customer updated from maintenance level 1 to ML 2. This left them booting to LED 554. We edited the ODM and it didn't really fix it.

A 3rd shift tech replaced the ODM with his pre-upgrade version. His package database was now massively out of date and installations failed.

Another 3rd shift tech convinced him to restore his mksysb and try again.

Later, this tech told the person that a mismatched mksysb, lppsource and SPOT were OK.

The next update was that the customer's database was failing with lots of libc.a errors about unexported symbols in /unix.

Then yet another tech tells the customer it's an application problem.

Complaint
The last tech is the guy who has kick ass numbers and they love him. He gets more in raises than I do, he comes in late, leaves early, and the reason he works more problems is because he doesn't actually put any effort into working them, and he closes them before they are resolved, because customers can call in and reopen them. Then, they get someone else, and then, when they are surveyed, the LAST person to work on them gets the ding.

Like me. I got a crappy rating because someone complained that it took them a month to get to me who solved their problem in 15 minutes, and because I didn't take as many calls in the off hours as the prime-shift people do.

I tried discussing why this doesn't make sense to 2 managers and they were all trying to convince me that a 3 really is OK. One said it was based on the amount of money available for raises and bonuses and they just couldn't fit in another 2.

It's been months and I'm still pissed, so to me, that means that it wasn't properly justified.

Plus, I was told, do such and such to get awards. Some of the things just couldn't be done. I tried. How do you create a database when you are not allowed to run your own server, and when you ask for the server admins to do it, your request is ignored and sits there for months?

And training? I give people pre-digested information, and they don't read it. It's only training if I put them into a conference room and force feed it to them. But, I'm not told this until AFTER review time.

And the Web page. I clean it up, re-organize it, fix all the bad links and convert it to the new globally mandated style. I make three perl script tools as requested. "thanks."

I'm just bitching. I know "life is not fair" but when things don't make sense, it really torques me out of shape.

Oh, and along those lines, they finally did something about our "stats". When they remove the woman who is both new and on maternity leave, and they remove the 2 3rd shift people, even leaving weekends and the 6-9pm stretch in there, they find that we have been 68-73% utilized for our target of 54%. "Keep up the good work." Every 2 days we get a message, "Why are you all not available."

The whole concept of researching or working in the lab is pushed aside, but we're not allowed to escalate calls anywhere without approval, and we get alot of pushback on that too. Oh, but if you have calls that you miss a call back on, you're written up in a report also.

I don't enjoy my job anymore. I like most of the people I work with, though I'm not part of their social club. About half of the people have their happy hour and such, and it's only through "leaked information" that the rest of us have even known about it.

There's no team social lunch type things, which might help, so maybe that should be suggested.
Tags:


  • 1
(Deleted comment)
I was thinking about asking for the adv support position, but I wasn't sure I wanted to be stuck there for another couple of years. I will ask to move somewhere else when I figure out whether we should . From where I am now, I would want to be in BC or in a L2/L3 org somewhere.

(Deleted comment)
*nod* That's my biggest complaint. It's not enough to be just the same version/release. The same ML is usually good enough, though on far extremes of sub-ML patches there can be a problem.

Generally, I try to impress upon people to apply patches to the lppsource and spot, or make a new one, whenever systems get updated.

(Deleted comment)
Yah, it's been a pain since the pSeries HW came out basically, which includes the winterhawk and nighthawk nodes too I think.

(Deleted comment)
*nod* I think it's time to move on. I don't trust my management structure anymore, though at least our new manager is competant. She's not as on top of things as our past manager, but she's better than many. It helps that she was service planner for AIX support for some number of years first, but she's never actually worked in support.

  • 1