AP530ww suddenly having strange LSQL errors

[Robbert Jan van de Veld]

Hello all,

We have about 12 AP530 access points in three different buildings. One day all of them were unresponsive. I couldn't reach them by web or cli. When I looked at the AP all LED's were on except the power LED.

When I took them to my desk and hooked them up to my computer and I tried to show or change something I was seeing the following weird LSQL errors:

Error getting data via LSQL.
lsqlsk_eval connect: Connection refused
Error getting data via LSQL.
lsqlsk_eval connect: Connection refused
Error getting data via LSQL.
lsqlsk_eval connect: Connection refused
Error getting data via LSQL.
lsqlsk_eval connect: Connection refused
Error getting data via LSQL.
lsqlsk_eval connect: Connection refused
Error getting data via LSQL.

I'm also using ProCurve Manager 2.3+ and there was nothing weird in the event log. However in the device syslog I came across some unexplainable notifications that would happen maybe three times a minute:

5>Jan 1 00:09:00 __alloc_pages: 0-order allocation failed (gfp=0x1d2/0)
4>Jan 1 00:09:00 VM: killing process snmpd

When I disconnected the network cable I suddenly could change all the settings. I changed the IP from DHCP to a static one and connected the network cable again. This seemed to be a workaround for the weird LSQL errors and the sudden halts, but the access point itself was still was very slow or even unresponsive at some times.

Also when I changed this the AP's were very jumpy in ProCurve Manager, they would very frequently change to red or green.

I tried a return to factory defaults and started building from scratch, but this didn't work either the errors would return from the beginning.

I use software version WA.02.19 for all my access points. The strange thing about this all was that it happened on all AP's at the same time when no change was made.

I hope anyone can help me with this otherwise I have to send them back to HP for repair.

Greetings,

Robbert Jan