Fibre Channel (SAN)

Reply
it4
New Contributor
Posts: 3
Registered: ‎10-20-2009

FOS 7.4.1 memory leak in several Brocade 5100?

We have several 5100 Switches that are experiencing growth of memory consumption (memory leak?)

MEM growth is about +1% usage per day - from last firmware update about 4 weeks ago memory usage started growing all the time until we get MAPS warning at >75%

 

Now we have FOS 7.4.1 in all of them (updated regularly 2..4- times a year, have gone through 7.2, 7.3 etc)

 

With earlier FOS versions (earlier version was 7.4.0b) we did not have any problems with memory consumpition growth

 

Installed FOS is "an IBM version" from/through IBM support site - I do not know if there is any difference with the "regular Brocade FOS"?

 

TOP (memory) shows nothing particular in any of the switches - eg. here: (this one is about 65% mem now but growing every day)

 

top - 08:44:30 up 27 days, 16:55,  1 user,  load average: 0.52, 0.75, 0.66
Tasks: 115 total,   1 running, 114 sleeping,   0 stopped,   0 zombie
Cpu(s):  1.6%us,  0.3%sy,  0.0%ni, 98.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:    504244k total,   491604k used,    12640k free,    13644k buffers
Swap:        0k total,        0k used,        0k free,   156880k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
 1447 root      16   0  143m  56m  44m S  0.0 11.5 202:18.28 iswitchd
 1411 root      16   0  135m  18m 9.8m S  0.0  3.7   1023:00 mdd0
 1413 root      16   0  149m  17m  11m S  0.0  3.5 166:22.11 cald0
 1449 root      16   0  112m  16m  12m S  0.0  3.3   0:59.54 nsd0
 1508 root      16   0 91236  14m  12m S  0.0  3.0   1438:38 0.weblinker.fcg
 1330 root      16   0 48140  14m 9488 S  0.0  2.9   0:00.60 traced
 1463 root      16   0  129m  13m  10m S  0.0  2.7 536:08.46 snmpd
 1440 root      16   0 99692  10m 6044 S  0.0  2.2  21:00.97 fspfd0
 1432 root       6 -10  106m  10m 7732 S  0.0  2.1 127:51.55 fabricd0
 1414 root      16   0  103m 9532 6444 S  0.0  1.9   1:34.06 essd0
 1457 root      16   0  107m 9324 7336 S  0.0  1.8   0:00.65 secd0
 1467 root      16   0  108m 9072 6420 S  0.0  1.8   0:04.61 zoned0
 1328 root      16   0 86060 8636 7408 S  0.0  1.7   0:04.96 raslogd
 1439 root      16   0 97368 8600 6172 S  0.0  1.7   0:00.80 ficud0
 1415 root      16   0 98208 8480 6448 S  0.0  1.7   0:00.35 evmd0
 1456 root      16   0 74276 8420 6588 S  0.0  1.7   0:00.80 rpcd0
 1465 root      16   0  105m 8204 5192 S  0.0  1.6  70:38.17 trafd0
 1448 root      16   0  108m 8160 6032 S  0.0  1.6   2:06.62 msd0
 1332 root      16   0  124m 8064 5832 S  0.0  1.6 318:46.30 emd0
 1260 root      16   0 83836 7908 5948 S  0.0  1.6 158:36.86 hasmd0
 1433 root      16   0 72852 7612 4984 S  0.0  1.5  24:08.84 fcpd0
 1450 root      16   0 72692 7568 5600 S  0.0  1.5  50:32.99 pdmd0
 1438 root      16   0 95912 7456 5168 S  0.0  1.5   1:28.68 fdmid0
 1479 root      16   0 63576 7428 5292 S  0.0  1.5   0:57.80 npd
 1320 root      16   0 68128 7376 5120 S  0.0  1.5   0:00.56 pdmd0
 1410 root      16   0  100m 7376 5308 S  0.7  1.5  61:31.52 webd0
 1455 root      16   0 78920 7348 5308 S  0.0  1.5   0:00.40 psd0
 1441 root      16   0  117m 7216 5120 S  0.0  1.4   0:00.28 rcsd0
 1458 root      16   0  102m 7088 5140 S  0.0  1.4  35:51.76 authd0
 1466 root      16   0 71960 7008 5216 S  0.0  1.4   4:14.70 tsd0
 1412 root      16   0 70356 6956 5136 S  0.0  1.4   0:01.10 arrd0
 1518 root      16   0 70744 6936 4984 S  0.0  1.4   0:00.24 agd0
13563 root      16   0 48100 6864 5712 S  0.3  1.4   0:00.88 sshd
 1478 root      16   0 86760 6748 4924 S  0.0  1.3   1:10.51 vsd0
 1253 root      16   0 46464 6484 4848 S  0.0  1.3   0:00.81 superd
 1334 root      18   0 22216 6304 4572 S  0.0  1.3   0:00.76 porttestd
 1331 root      16   0 49740 6108 4436 S  0.0  1.2   0:01.07 diagd0
 1321 root      16   0 52660 5676 4052 S  0.0  1.1   0:00.49 lfmd

 

Currently we clear the memory overgrowth situation by issuing first:

supportinfoclear --clear -RASlog -force (this is a precaution only)

 

and then:

 

hareboot

 

=> memory consuption goes down to about 45-47% level - but starts growing again

 

Has anyone else experienced similar symptoms?

 

Any advice on how to debug the problem?

 

External Moderator
Posts: 5,034
Registered: ‎02-23-2004

Re: FOS 7.4.1 memory leak in several Brocade 5100?

Is a Defect.

 

Brocade non–director class switches (see the Products Affected section for the complete list) running
FOS versions v7.2.1f, v7.3.1c or v7.4.1 are at risk for this problem.

 

will be fixed whit 7.4.1a

TechHelp24
Contributor
Posts: 36
Registered: ‎07-19-2007

Re: FOS 7.4.1 memory leak in several Brocade 5100?

Can I ask you to clarify that just a bit:

 

Are you saying that version 7.4.0a has neither the defect or the resolution?

it4
New Contributor
Posts: 3
Registered: ‎10-20-2009

Re: FOS 7.4.1 memory leak in several Brocade 5100?

We ran 7.4.0a and b for quite some time without problems. So my quess is that those do not have the same problem.

it4
New Contributor
Posts: 3
Registered: ‎10-20-2009

Re: FOS 7.4.1 memory leak in several Brocade 5100?

Many thanks!

Join the Community

Get quick and easy access to valuable resource designed to help you manage your Brocade Network.

vADC is now Pulse Secure
Download FREE NVMe eBook