1 |
On 05/02/2013 02:47 PM, Alan Mackenzie wrote: |
2 |
> On Thu, May 02, 2013 at 01:15:58PM -0400, Michael Mol wrote: |
3 |
>> On 05/02/2013 12:58 PM, Alan Mackenzie wrote: |
4 |
>>> On Thu, May 02, 2013 at 12:33:37PM -0400, Michael Mol wrote: |
5 |
>>>> On 05/02/2013 12:27 PM, Alan Mackenzie wrote: |
6 |
>>>>> Hi, Gentoo. |
7 |
> |
8 |
>>>>> I've just built libreoffice-3.6.6.2 and it took 2 hours 10 |
9 |
>>>>> minutes on my 2.6 GHz quad core Athlon 2. It used to take about |
10 |
>>>>> an hour. |
11 |
> |
12 |
>>>>> Watching the build, it became evident that the first 50 minutes |
13 |
>>>>> or so was taken up by several hundred mkdir operations (more |
14 |
>>>>> precisely, mkdir -p <long path>). Some of these mkdir's would |
15 |
>>>>> take, perhaps, a minute to execute. All the while, top showed |
16 |
>>>>> make taking 100% of one core. |
17 |
> |
18 |
>>>>> There seems to be something suboptimal here. Has anybody else |
19 |
>>>>> seen this, or does anybody have any ideas how to fix the |
20 |
>>>>> problem? |
21 |
> |
22 |
>>>> Long delays suggest a timeout of some sort. |
23 |
> |
24 |
>>> OK. As a matter of interest, some of the mkdirs executed relatively |
25 |
>>> quickly - perhaps in 0.5 seconds. I never saw the screen whizzing by |
26 |
>>> as I ought to have done, though. |
27 |
> |
28 |
>> Hm. |
29 |
> |
30 |
> |
31 |
>>>> First thing I'd look at is the filesystem underneath, and the disk |
32 |
>>>> underneath that. |
33 |
> |
34 |
>>> My /var is an ext3 LVM partition, doubled up on a RAID-1 disk array. |
35 |
> |
36 |
>> How full is the ext3 partition? What options do you have enabled on it? |
37 |
>> (e.g. dir indexing?) |
38 |
> |
39 |
> root@acm ~ # df /var |
40 |
> Filesystem 1K-blocks Used Available Use% Mounted on |
41 |
> /dev/mapper/vg-var 12385456 1959860 9796580 17% /var |
42 |
> |
43 |
>>> In the middle of the mkdiring, I checked there were enough inodes |
44 |
>>> free (there were). I've no reason to suspect the disk drives might |
45 |
>>> be flaky. |
46 |
> |
47 |
>> Well, you kinda do, now; |
48 |
> |
49 |
> The reason I say this is that building the last ?one/two/three versions |
50 |
> of libreoffice also gave me this grief, but I haven't noticed anything |
51 |
> else amiss. |
52 |
> |
53 |
>> you have evidence that at least some disk access is unusually slow. |
54 |
>> Check dmesg for disk I/O errors (unlikely to be reported at this point; |
55 |
> |
56 |
> Nothing awry in dmesg. |
57 |
> |
58 |
>> I'm sure you checked whether your RAID was in a degraded state), |
59 |
> |
60 |
> cat /proc/mdstat shows everything in order. |
61 |
> |
62 |
>> and run commanded smartctl tests on the disks. |
63 |
> |
64 |
> That I haven't done, yet. |
65 |
> |
66 |
>>>> Second thing I'd look at is to see if permissions checks might be |
67 |
>>>> bouncing through something like kerberos, samba or ldap. Do you |
68 |
>>>> have any single-signon things configured on that machine? |
69 |
> |
70 |
>>> I've not got kerberos or samba installed. I appear to have ldap |
71 |
>>> (whatever that might be ;-). ls -lurt /usr/bin/ldap* shows these |
72 |
>>> binaries were last accessed (?used) on 2012-03-14. |
73 |
> |
74 |
>> It would be more a question of whether they were tied into PAM. |
75 |
> |
76 |
> OK. I'm sadly ignorant about PAM. :-( |
77 |
|
78 |
If you've just got a single box, it's very unlikely this is your problem. |
79 |
|
80 |
> |
81 |
>>> What exactly do you mean by "single-signon"? |
82 |
> |
83 |
>> Well, that was a slip of the tongue. More "central auth". I was |
84 |
>> wondering if there were any features installed on your system that are |
85 |
>> designed to check authorization against a server somewhere. (i.e. you |
86 |
>> can use an LDAP directory to centrally manage things like users, groups, |
87 |
>> etc.) |
88 |
> |
89 |
> Not that I know of. My machine is a mere desktop connected via a |
90 |
> router/modem to the net. I'd have no reason to install any auth stuff. |
91 |
> |
92 |
>> Technically, single-signon combines authorization checks with persistent |
93 |
>> authentication checks. Examples of this include kerberos, web session |
94 |
>> cookies and some uses of OAuth; once you're authenticated, the mechanism |
95 |
>> ensures you don't need to authenticate to another server in the same |
96 |
>> auth realm so long as your existing session hasn't expired. But this is |
97 |
>> less likely to be related to your problem than something seeking to ask |
98 |
>> a server if you have authorization to access something. |
99 |
> |
100 |
> If this were the case, what would libreoffice's build need to ask that no |
101 |
> other package stumbles over? |
102 |
> |
103 |
|
104 |
My presumption there was that this was a very recent thing, and LO's |
105 |
time-to-build makes it easier to observe. |
106 |
|
107 |
Anyway, floor's open to anyone else who might have an idea. |