☰

WAS 未处理请求因此网络不可用

tag-icon • tag-icon web-server java web-applications apache2 websphere
WAS 未处理请求因此网络不可用

我是 WAS（Websphere 应用服务器）支持新手，目前我的 WAS 出现了问题，我的 WAS 安装在 AIX 的 2 个服务器/节点下。

在调查时，我在我们的应用程序日志中发现有一些活动“执行缓存维护”：：

2017-01-14 01:31:52,619: [Cache Maintenance] com.ibm.srm.util.db.ServerCache refreshed
2017-01-14 01:31:53,314: [Cache Maintenance] Memory: available=[6884mb] used=[9500mb] %used avail=[58%] max=[16384mb] %used max=[58%] total=[16384mb] free=[6884mb] used by doMaintenance=[-251,201,3
92bytes] Time=[22,818ms]
2017-01-14 01:51:53,325:  -------- Performing Cache Maintenance --------
2017-01-14 01:51:53,325: null    : QN=319 Select * from perform.cache_timestamps where row_class_name not like '%Cache' and row_class_name not like '%(SRM 6.0)'
2017-01-14 01:51:53,333: Returning 19 data records, QN=319,2 columns, Time: 8ms  conn/query time: 5ms
2017-01-14 01:51:53,333: [Cache Maintenance] Memory: available=[5492mb] used=[10892mb] %used avail=[66%] max=[16384mb] %used max=[66%] total=[16384mb] free=[5492mb] used by doMaintenance=[532kb] Time=[8ms]

触发此活动后，我发现 mpmstats 中“bsy”的值不断增加，直到达到 MaxClient 的最大值“4000”::

[Sat Jan 14 01:38:58 2017] [notice] mpmstats: rdy 166 bsy 234 rd 0 wr 234 ka 0 log 0 dns 0 cls 0
[Sat Jan 14 01:38:58 2017] [notice] mpmstats: bsy: 234 in mod_was_ap22_http.c
[Sat Jan 14 01:48:58 2017] [notice] mpmstats: rdy 195 bsy 505 rd 0 wr 505 ka 0 log 0 dns 0 cls 0
[Sat Jan 14 01:48:58 2017] [notice] mpmstats: bsy: 505 in mod_was_ap22_http.c
[Sat Jan 14 01:58:58 2017] [notice] mpmstats: rdy 180 bsy 720 rd 0 wr 720 ka 0 log 0 dns 0 cls 0
[Sat Jan 14 01:58:58 2017] [notice] mpmstats: bsy: 720 in mod_was_ap22_http.c
[Sat Jan 14 02:08:59 2017] [notice] mpmstats: rdy 105 bsy 895 rd 1 wr 894 ka 0 log 0 dns 0 cls 0
[Sat Jan 14 02:08:59 2017] [notice] mpmstats: bsy: 894 in mod_was_ap22_http.c
[Sat Jan 14 02:18:59 2017] [notice] mpmstats: rdy 112 bsy 1088 rd 1 wr 1087 ka 0 log 0 dns 0 cls 0
[Sat Jan 14 02:18:59 2017] [notice] mpmstats: bsy: 1087 in mod_was_ap22_http.c
[Sat Jan 14 02:28:59 2017] [notice] mpmstats: rdy 158 bsy 1242 rd 1 wr 1241 ka 0 log 0 dns 0 cls 0

[..]

[Sat Jan 14 04:55:34 2017] [notice] mpmstats: rdy 0 bsy 4000 rd 0 wr 4000 ka 0 log 0 dns 0 cls 0
[Sat Jan 14 04:55:34 2017] [notice] mpmstats: bsy: 4000 in mod_was_ap22_http.c
[Sat Jan 14 04:57:04 2017] [notice] mpmstats: reached MaxClients (4000/4000)
[Sat Jan 14 04:57:04 2017] [notice] mpmstats: rdy 0 bsy 4000 rd 0 wr 4000 ka 0 log 0 dns 0 cls 0
[Sat Jan 14 04:57:04 2017] [notice] mpmstats: bsy: 4000 in mod_was_ap22_http.c
[Sat Jan 14 04:58:34 2017] [notice] mpmstats: reached MaxClients (4000/4000)
[Sat Jan 14 04:58:34 2017] [notice] mpmstats: rdy 0 bsy 4000 rd 0 wr 4000 ka 0 log 0 dns 0 cls 0
[Sat Jan 14 04:58:34 2017] [notice] mpmstats: bsy: 4000 in mod_was_ap22_http.c

看起来 WAS 直到达到最大值才会处理客户端请求。

问题是：

  1. 是否有任何日志可以让我检查为什么 WAS 直到达到最大值才处理客户端请求？

  2. “缓存维护”活动是否会阻止 WAS 处理客户端请求？因为正如我们的开发人员所说，此活动不应该导致此问题。

  3. 我可以采取什么程序来识别/解决这个问题？

答案1

第一步是在缓存维护期间和之后不久获取 javacore。它可能会告诉您哪些资源被滥用导致应用程序服务器基本上锁定。然后将其交给 com.ibm.srm 应用程序的适当支持团队。

相关内容