Did the computer id change for the system? There isn’t anything like deep freeze on the system is there?
No Sir. Nothing like that.
[thread resurrection]
I’m still seeing this daily with 9.5.12 clients. Mostly on macOS, but sometimes Windows too.
Was a PMR or RFE ever filed for this? (I know I didn’t!)
I know of one case that is currently open but it has been very hard to catch a machine “in the act”. Client is 9.5.10 and 9.5.11 in this case though. Did this behavior get worse with 9.5.12? Case number is TS001152528 in case anyone needs to reference it.
1 Like
It’s been around for a very long time. I think it might have gotten more frequent on 9.5.12 but I can’t be sure. I’m definitely used to seeing it, and have even developed fixlets to chuck computers back in the MSP bin they should be.
We have an open case and are pursuing it actively. We have managed to obtain debug and system and client logs from affected machines “in the act” of losing all their client settings, and have provided them to support. No conclusions to report yet, but we’ll keep the thread updated with developments.
In fact, right as I mentioned this, we heard back from some initial log analysis from the support engineer working on our ticket. It looks like the loss of client settings may stem from hard/non-graceful shutdown of the system. We’re going to do some additional debugging and see if we can gather more information.
[July 2019]
We’re still seeing this happening on 9.5.13 clients. It’s gotten quite tedious, and I’m close to putting in policy actions to deal with the customer areas that are easy to identify. (Lame but whatchugonnado)
Any news on isolating the cause, and hopefully a permanent fix?
-Andrew
Still working through this with support. We have an alternate binary to test out, but because of the sporadic nature of the failures, testing anything at scale is challenging. Will update the thread as we learn more.
It appears that we’ve got a custom build that really does fix the issue - we have deployed it to about 10,000 machines and none of them have lost client settings since. No word yet as to what the fix was, but one hopes that it will make it into a real release in the near future.
3 Likes
That’s great to hear! Thanks for all the time you (and everyone involved) has spent on this!
WOWOWOW!
Fantastic news! Thanks for all the work chasing it down!
Excellent… This has been a Pain in the proverbial…
Sorry to pull reso this, jtaven, has the fix been holding for you? Any word from our HCL friends that this could possibly make it into the next point release?
So far so good on the fix front. I haven’t pulled the numbers, but the last time our team checked, none of the patched systems had lost settings. Support indicated that the fix should make it in to the next point release, I believe, but I haven’t heard about when that might be.
Super excellent news! I’d love to know root cause of course, if only to dispell some of the wild theories we’ve been coming up with for it ourselves. At one point we working with IBM on a ticket for the issue, but ended up having to drop it when we just couldn’t capture the Mac system.logs they were looking for. It was never a system nearby of course… Thanks for keeping at em.
What version of BigFix is your alternate binary for? We are running 9.5.12 in production and appear to have a similar issue. Our lab environment has 9.5.14 and I’m not sure if it is fixed there or we just haven’t encountered the bug yet.
9.2.13 addressed this issue:
Issue BFP-13961 - APAR KB0069026 - MAC: Client settings may disappear from BESAgent plist file
see: https://support.bigfix.com/bes/changes/fullchangelist-95.txt
1 Like
I hate to resurrect this topic again, but I just found a glitch in my reporting and it looks like the problem is present again now that we have upgraded to 10.0.0.133. I’m starting back up working with HCL to address this, but people with Mac fleets stay tuned.