[Tanster]

Here’s what I did (whether it’s the right/best way or not, it seems to work) using just the GUI:

1. Turned on LACP bonding in the gigabit switch (ZyXEL GS-2024) for the two ports.

2. Fired up the 2nd NIC via the GUI and ARD (for everything, no CLI involved) and gave it a static IP.

3. Disconnected from 1st NIC and reconnected via 2nd NIC.

4. Created a 1 NIC LACP bonded pipe with the 1st NIC and configured it with the same static IP it had before as a singleton.

5. Disconnected from the 2nd NIC and reconnected via 1st NIC.

6. Added the 2nd NIC into the LACP bonded pipe created in #4 above by just editing it in the Network->Network config pane.

7. Checked using “changeip -checkhostname” to make sure nothing was upsetting to the Xserve.

Voila. Didn’t end up pulling out the rug from under myself either. And the LACP bonding seems to work even after a restart.

[Tanster]

[QUOTE][u]Quote by: Moofo[/u][p][b]Just so everyone knows.

The issue is really solved on the Intel OS X server as far as I know. The studio where the Xserve is being used is doing lots of search and everything seems to be fine now. In fact, I had to turn off the machine one time: to clean the vents ! [/b]

[/p][/QUOTE]

Unfortunately, we have a Xserve G5 at 10.4.9 so that still leaves us out in the cold…

In desperation, we even tried this “fix” in [b]http://www.creativetechs.com/iq/network_crashes_in_cs2_use_adobe_dialog.html[/b] to mixed results–some workstations still hang, some still kernel panic but at least AFP on the Xserve G5 doesn’t max out any more. And the Xserve G5 still doesn’t indicate in any of its logs what’s happening aside from noting that the client workstation isn’t responding in the AFP access log: “Client no response …”

At this point, I’m seriously considering the witch doctor masks and sacrificial chickens/goats/whatever… 🙄

[Tanster]

I’ve been watching this thread as I have an Xserve G5 behaving in exactly the same way. Unfortunately, it’s already been upgraded to 10.4.9 and it’s still exhibiting the same issues with “CatSearch starting…” affecting AFP access. Anybody else still seeing this issue with 10.4.9 or has it finally been eradicated with PPC machines and my Xserve is just an anomaly?

Also, would going through the entire data volumes where searches are likely to be performed and deleting the .DS_Store help as a temporary stopgap measure?

[Author]

Posts