Texis on UNIX - error log

[skalyanaraman]

Hi,
We kinda experienced a slowness in searching the texis databases on our UNIX boxes this morning. We had no clue. But the /var/adm/messages log file has a bunch of warning for the monitor program right about at the same time. Our searching is back to normal now. So, could you please take a look at the following entries and let us know if this problem is serious and also how to fix it. Here is the log listing...

Jul 2 06:52:06 texisup2 genunix: [ID 470503 kern.warning] WARNING: Sorry, no swap space to grow stack for pid 22740 (monitor)
Jul 2 06:52:06 texisup2 last message repeated 7 times
Jul 2 07:00:41 texisup2 genunix: [ID 470503 kern.warning] WARNING: Sorry, no swap space to grow stack for pid 22800 (monitor)
Jul 2 07:00:41 texisup2 last message repeated 8 times
Jul 2 07:47:07 texisup2 genunix: [ID 470503 kern.warning] WARNING: Sorry, no swap space to grow stack for pid 23627 (monitor)
Jul 2 07:47:07 texisup2 last message repeated 8 times
Jul 2 08:28:39 texisup2 last message repeated 2 times
Jul 2 08:42:51 texisup2 genunix: [ID 470503 kern.warning] WARNING: Sorry, no swap space to grow stack for pid 25360 (monitor)
Jul 2 08:42:51 texisup2 last message repeated 8 times


Thanks!!

[Kai]

Those monitor processes might have been trying to update a Metamorph index, which can consume a lot of memory and hence swap space. Make sure you've got plenty of swap, generally at least as much as physical RAM.

Also, try updating the Metamorph indexes manually, with tsql or texis: you can set indexmem to a lower value to reduce the memory usage of index updating (with little or no impact on speed, though temp disk usage may increase). Eg. "set indexmem=20" in SQL just before the create metamorph statement.