| 1 |
Steve wrote: |
| 2 |
> This is very frustrating... having played around, the syslog-ng tends |
| 3 |
> towards using 100% CPU when my server is otherwise quiet - if, and |
| 4 |
> only if, I have the program destination... even if the destination is |
| 5 |
> not used. |
| 6 |
|
| 7 |
Oh, and strace shows syslog-ng frantically polling file-descriptor 3... |
| 8 |
which, I presume, is the pipe to the sshguard process. |
| 9 |
|
| 10 |
> poll([{fd=6, events=0}, {fd=4, events=POLLIN}, {fd=10, events=POLLIN}, |
| 11 |
> {fd=3, events=POLLIN}, {fd=13, events=POLLIN}, {fd=14, events=POLLIN}, |
| 12 |
> {fd=7, events=POLLIN}, {fd=8, events=POLLIN}], 8, 1178000) = 1 |
| 13 |
> ([{fd=6, revents=POLLERR}]) |
| 14 |
> gettimeofday({1237836567, 385148}, NULL) = 0 |
| 15 |
> gettimeofday({1237836567, 385178}, NULL) = 0 |
| 16 |
> poll([{fd=6, events=0}, {fd=4, events=POLLIN}, {fd=10, events=POLLIN}, |
| 17 |
> {fd=3, events=POLLIN}, {fd=13, events=POLLIN}, {fd=14, events=POLLIN}, |
| 18 |
> {fd=7, events=POLLIN}, {fd=8, events=POLLIN}], 8, 1178000) = 1 |
| 19 |
> ([{fd=6, revents=POLLERR}]) |
| 20 |
> gettimeofday({1237836567, 385506}, NULL) = 0 |
| 21 |
> gettimeofday({1237836567, 385712}, NULL) = 0 |
| 22 |
> poll([{fd=6, events=0}, {fd=4, events=POLLIN}, {fd=10, events=POLLIN}, |
| 23 |
> {fd=3, events=POLLIN}, {fd=13, events=POLLIN}, {fd=14, events=POLLIN}, |
| 24 |
> {fd=7, events=POLLIN}, {fd=8, events=POLLIN}], 8, 1178000) = 1 |
| 25 |
> ([{fd=6, revents=POLLERR}]) |
| 26 |
> gettimeofday({1237836567, 386027}, NULL) = 0 |
| 27 |
> gettimeofday({1237836567, 386241}, NULL) = 0 |
| 28 |
Do others get this behaviour - is this a bug in syslog-ng? |
Archives